diff --git a/.gitattributes b/.gitattributes
index a547963b2f6f5092906b359d9ed0c913c8b6898e..9f684ccc05817387cac19b4d973ae3d9dd0b4081 100644
--- a/.gitattributes
+++ b/.gitattributes
@@ -10,3 +10,5 @@
*.avi filter=lfs diff=lfs merge=lfs -text
*.mov filter=lfs diff=lfs merge=lfs -text
*.gif filter=lfs diff=lfs merge=lfs -text
+*.png filter=lfs diff=lfs merge=lfs -text
+*.jpg filter=lfs diff=lfs merge=lfs -text
diff --git a/README.md b/README.md
index 44632f513f93d08f01e65c62d9bc25fed7a697de..8e91a03aec4c0633d9946ac7665b83811217dbdc 100644
--- a/README.md
+++ b/README.md
@@ -1,170 +1,14 @@
-# fast_dwpose
+# For developers
-The goal of Easy DWPose is to provide a generic, reliable, and easy-to-use interface for making skeletons for ControlNet.
+## tutorial00 Use DWpose to preprocess the original sign language video
-SF do some improve for [easy-dwpose](https://github.com/reallyigor/easy_dwpose), named it fast-dwpose.
+## tutorial01 prepare pose dict
-## Installation
+## tutorial02 get pose video from text input
-### PIP
+## tutorial03 render pose video
-```bash
-pip install easy-dwpose
-```
-## Quickstart
+# For Users
-### In you own .py scrip or in Jupyter
-
-```python
-import torch
-from PIL import Image
-import numpy as np
-import json
-
-from easy_dwpose import DWposeDetector
-
-#####---------Setup init
-# You can use a different GPU, e.g. "cuda:1"
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
-detector = DWposeDetector(device=device)
-input_image = Image.open("assets/pose.png").convert("RGB")
-
-
-#####---------Get both the skeleton image
-# SF: skeleton should be a kind of img
-skeleton = detector(input_image, output_type="pil", include_hands=True, include_face=True)
-# Save the skeleton image
-skeleton.save("skeleton.png")
-
-
-#####---------Get pose data
-# SF: pose_data should be numpy/tensor
-# # This returns the dictionary
-pose_data = detector(input_image, draw_pose=False)
-
-# Save the skeleton pose information:
-# Option 1: Save as NPY file
-np.save('pose_data.npy', pose_data)
-
-# Option 2: Save as NPZ file
-np.savez('pose_data.npz', **pose_data)
-
-# Option 3: Save as JSON file
-# Convert numpy arrays to lists for JSON serialization
-pose_data_json = {k: v.tolist() if isinstance(v, np.ndarray) else v for k, v in pose_data.items()}
-with open('pose_data.json', 'w') as f:
- json.dump(pose_data_json, f)
-```
-
-
-
- | Input |
- Output |
-
-
-
-
- 
- |
-
-
- 
- |
-
-
-
-### On a video
-
-```bash
-python scripts/inference_on_video.py --input assets/dance.mp4 --output_path result.mp4
-```
-
-
-
- | Input |
- Output |
-
-
-
-
- 
- |
-
-
- 
- |
-
-
-
-### On a folder of images
-
-```bash
-python scripts/inference_on_folder.py --input assets/ --output_path results/
-```
-
-### Easy-DWPose Custom skeleton drawing
-
-By default, we use standart skeleton drawing function but several projects change it (e.g. [MusePose](https://github.com/TMElyralab/MusePose)). Modify it or write your own from scratch!
-
-```python
-from PIL import Image
-from easy_dwpose import DWposeDetector
-from easy_dwpose.draw.musepose import draw_pose as draw_pose_musepose
-
-detector = DWposeDetector(device="cpu")
-input_image = Image.open("assets/pose.png").convert("RGB")
-
-skeleton = detector(input_image, output_type="pil", draw_pose=draw_pose_musepose, draw_face=False)
-skeleton.save("skeleton.png")
-```
-
-### SF Custom skeleton drawing
-
-I prefer ControlNext style, I have developed a visualization method and placed it in `./easy_dwpose/draw/controlnext.py`. It does not integrate with easy dwpose and the calling method is slightly different:
-
-```python
-import torch
-from PIL import Image
-import numpy as np
-import json
-
-from easy_dwpose import DWposeDetector
-from easy_dwpose.draw.controlnext import draw_pose, process_pose_data
-
-#####---------Setup init
-# You can use a different GPU, e.g. "cuda:1"
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
-detector = DWposeDetector(device=device)
-input_image = Image.open("assets/pose.png").convert("RGB")
-
-#####---------Custom ControlNext drawing style
-# Get pose data for custom drawing
-pose_data = detector(input_image, draw_pose=False)
-
-# Get image dimensions
-width, height = input_image.size
-
-# Process the pose data for custom drawing
-processed_pred = process_pose_data(pose_data, height, width)
-
-# Draw pose using custom ControlNext style
-vis_img = draw_pose(
- pose=processed_pred,
- H=height,
- W=width,
- include_body=True,
- include_hand=True,
- include_face=True
-)
-
-# Convert to PIL Image and save (vis_img is in CHW format)
-custom_skeleton = Image.fromarray(vis_img.transpose(1, 2, 0))
-custom_skeleton.save("skeleton_controlnext.png")
-```
-
-## Acknowledgement
-
-We thank the original authors of the [DWPose](https://github.com/IDEA-Research/DWPose) for their incredible models!
-
-Thanks for open-sourcing!
+## start00 just start as a user
\ No newline at end of file
diff --git a/assets/WLASL_01214/crop_frame/00000001.jpg b/assets/WLASL_01214/crop_frame/00000001.jpg
deleted file mode 100644
index fba30ec43d2c38aa9b39cbec0185da5e39d75ce1..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000001.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000002.jpg b/assets/WLASL_01214/crop_frame/00000002.jpg
deleted file mode 100644
index 858ca3a360e344ef2cafb5a5fce448cd2c376ecd..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000002.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000003.jpg b/assets/WLASL_01214/crop_frame/00000003.jpg
deleted file mode 100644
index 2d7eeedc41a5476799155be539aa0466268678ee..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000003.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000004.jpg b/assets/WLASL_01214/crop_frame/00000004.jpg
deleted file mode 100644
index a11e39d352e68f488cc9ef5ee9d6abc62f0b3fdc..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000004.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000005.jpg b/assets/WLASL_01214/crop_frame/00000005.jpg
deleted file mode 100644
index 73db113caeef196ddfaa0198a8e04387344ef6e7..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000005.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000006.jpg b/assets/WLASL_01214/crop_frame/00000006.jpg
deleted file mode 100644
index 06a4746f27bbcfff4225bd9241b0b6350f8edeec..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000006.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000007.jpg b/assets/WLASL_01214/crop_frame/00000007.jpg
deleted file mode 100644
index d35f2ad3b1346f05308e0f802e5be2ca7b6e4448..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000007.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000008.jpg b/assets/WLASL_01214/crop_frame/00000008.jpg
deleted file mode 100644
index a6202a3691db41cd42485c7d4833e819b7f46f2b..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000008.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000009.jpg b/assets/WLASL_01214/crop_frame/00000009.jpg
deleted file mode 100644
index e2fa902e9b7435e993497025b6c8f9220e562310..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000009.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000010.jpg b/assets/WLASL_01214/crop_frame/00000010.jpg
deleted file mode 100644
index 7f6c29a812ae5e943aeb93eec15aa4f4cb4412c9..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000010.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000011.jpg b/assets/WLASL_01214/crop_frame/00000011.jpg
deleted file mode 100644
index 5ad6ce3e0df36c258c5739133f4eb8d912551fdb..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000011.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000012.jpg b/assets/WLASL_01214/crop_frame/00000012.jpg
deleted file mode 100644
index 9b9794c9487fc95ecc7e7c2971df463b09766d94..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000012.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000013.jpg b/assets/WLASL_01214/crop_frame/00000013.jpg
deleted file mode 100644
index 02f5222a3ffae36f2a0d8ec75f776784b78bd98c..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000013.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000014.jpg b/assets/WLASL_01214/crop_frame/00000014.jpg
deleted file mode 100644
index b0169d98a8baa6f10bb8618f1d696f0fb97f94c4..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000014.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000015.jpg b/assets/WLASL_01214/crop_frame/00000015.jpg
deleted file mode 100644
index dadedd3da4bc5a34e255e0135edd5bbce6fd3c78..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000015.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000016.jpg b/assets/WLASL_01214/crop_frame/00000016.jpg
deleted file mode 100644
index f61107369174f37dc44df3fc98aaef11c99c75a7..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000016.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000017.jpg b/assets/WLASL_01214/crop_frame/00000017.jpg
deleted file mode 100644
index 9695cb8ddad0cc383407ca5466bcde990a3e3885..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000017.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000018.jpg b/assets/WLASL_01214/crop_frame/00000018.jpg
deleted file mode 100644
index 76bdf7b8f97d995815f430eb5d9702ed78d2b722..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000018.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000019.jpg b/assets/WLASL_01214/crop_frame/00000019.jpg
deleted file mode 100644
index 48634d6733cbcba849ae9746dfa4f517f1c2e579..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000019.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000020.jpg b/assets/WLASL_01214/crop_frame/00000020.jpg
deleted file mode 100644
index 539627b24007515f82cbb8c909aaebcc021c69ed..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000020.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000021.jpg b/assets/WLASL_01214/crop_frame/00000021.jpg
deleted file mode 100644
index 745b8b2fd143b75a23d39550007f2bf55748d391..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000021.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000022.jpg b/assets/WLASL_01214/crop_frame/00000022.jpg
deleted file mode 100644
index e758b9d04fd6ee18ea4a1745d19742696777ba6a..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000022.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000023.jpg b/assets/WLASL_01214/crop_frame/00000023.jpg
deleted file mode 100644
index 3873edf4ab1e2bffffedb0b258a923adbfe8bf71..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000023.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000024.jpg b/assets/WLASL_01214/crop_frame/00000024.jpg
deleted file mode 100644
index d6d58c81dde92b282943f2c3dd417b66b16a6856..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000024.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000025.jpg b/assets/WLASL_01214/crop_frame/00000025.jpg
deleted file mode 100644
index d46af0f9b26dfe60d113d195bd8e9bde67c8732c..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000025.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000026.jpg b/assets/WLASL_01214/crop_frame/00000026.jpg
deleted file mode 100644
index 1b54d3d94ab5fc84be189a05fd32b9473190fbed..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000026.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000027.jpg b/assets/WLASL_01214/crop_frame/00000027.jpg
deleted file mode 100644
index 1784a7997c8ed6075e65e6bbdf8d99416dc88106..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000027.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000028.jpg b/assets/WLASL_01214/crop_frame/00000028.jpg
deleted file mode 100644
index bacba08c4a982ebb99a242d27c35e44c0774f537..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000028.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000029.jpg b/assets/WLASL_01214/crop_frame/00000029.jpg
deleted file mode 100644
index be8fc02480b9b9d41bc48ec6e88cfcb55d8c4dd3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000029.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000030.jpg b/assets/WLASL_01214/crop_frame/00000030.jpg
deleted file mode 100644
index 7be18a67d1a737b52ad1d9c332ca3501eae5ed6d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000030.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000031.jpg b/assets/WLASL_01214/crop_frame/00000031.jpg
deleted file mode 100644
index 92335ac90e790bbdc01ec1c98adcdbfeef33b627..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000031.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000032.jpg b/assets/WLASL_01214/crop_frame/00000032.jpg
deleted file mode 100644
index b1e7ba6b0c6d5790083e13c9e848ea9e6a342c0b..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000032.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000033.jpg b/assets/WLASL_01214/crop_frame/00000033.jpg
deleted file mode 100644
index 0072060e6137179f30adc614cda2e4f1dd91b156..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000033.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000034.jpg b/assets/WLASL_01214/crop_frame/00000034.jpg
deleted file mode 100644
index 47129eeca672d14c2de79049880ed7ecb541fd02..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000034.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000035.jpg b/assets/WLASL_01214/crop_frame/00000035.jpg
deleted file mode 100644
index 3e3f29d8be95b6274f29f600eb4c5fe996851f2d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000035.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000036.jpg b/assets/WLASL_01214/crop_frame/00000036.jpg
deleted file mode 100644
index 8ac057e3f0793dde524e8b12751abdeede1823db..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000036.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000037.jpg b/assets/WLASL_01214/crop_frame/00000037.jpg
deleted file mode 100644
index bbfaa8e3cec7209196befaca730fad7b39d64ec4..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000037.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000038.jpg b/assets/WLASL_01214/crop_frame/00000038.jpg
deleted file mode 100644
index 00ba15c138f1dd3b6da2276020659f90aca4f580..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000038.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000039.jpg b/assets/WLASL_01214/crop_frame/00000039.jpg
deleted file mode 100644
index 9becaf406fc4ce19f405c7737ad5b7fee62705f0..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000039.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000040.jpg b/assets/WLASL_01214/crop_frame/00000040.jpg
deleted file mode 100644
index 2db90bf7959a3e8e0c628b48185d2e7a12e05b81..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000040.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/crop_frame/00000041.jpg b/assets/WLASL_01214/crop_frame/00000041.jpg
deleted file mode 100644
index 696a731c649b21928a5017438ceca1a10f2aa601..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/crop_frame/00000041.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000001.png b/assets/WLASL_01214/dwpose_overlay_img/00000001.png
deleted file mode 100644
index a6d282b28420b9069223a961dede3b8609011233..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000001.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000002.png b/assets/WLASL_01214/dwpose_overlay_img/00000002.png
deleted file mode 100644
index 5417902cb564c879bf49e40ee93a1aa79e49318b..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000002.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000003.png b/assets/WLASL_01214/dwpose_overlay_img/00000003.png
deleted file mode 100644
index c2f4850e6d43534fc11c5b74afcab0f46643bb7d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000003.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000004.png b/assets/WLASL_01214/dwpose_overlay_img/00000004.png
deleted file mode 100644
index 87d1600130229c5b919a878de9c88d343deafe9f..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000004.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000005.png b/assets/WLASL_01214/dwpose_overlay_img/00000005.png
deleted file mode 100644
index 1e446e0519f4b1eaf78960eb92d2d516d9f73c0f..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000005.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000006.png b/assets/WLASL_01214/dwpose_overlay_img/00000006.png
deleted file mode 100644
index 5e271ee0e3f28ab851d0797e3c6f0ac9be27bed7..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000006.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000007.png b/assets/WLASL_01214/dwpose_overlay_img/00000007.png
deleted file mode 100644
index 531f8e3f7d2ae596f88d2187b7409c67e3ac9dc6..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000007.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000008.png b/assets/WLASL_01214/dwpose_overlay_img/00000008.png
deleted file mode 100644
index d679cce1528178844c0e3870273825ed592e84f3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000008.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000009.png b/assets/WLASL_01214/dwpose_overlay_img/00000009.png
deleted file mode 100644
index 7d082a1cde83d75ad8e674c2525464471a69d366..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000009.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000010.png b/assets/WLASL_01214/dwpose_overlay_img/00000010.png
deleted file mode 100644
index 2198a9f1795ab16971cb75a7ebcd8e32ab06c38e..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000010.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000011.png b/assets/WLASL_01214/dwpose_overlay_img/00000011.png
deleted file mode 100644
index d12951dd7e4086efcfa1740c97f7177a6d00b7cd..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000011.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000012.png b/assets/WLASL_01214/dwpose_overlay_img/00000012.png
deleted file mode 100644
index c4bc6735ee9461199437cca8c9693ff9bf6414c0..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000012.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000013.png b/assets/WLASL_01214/dwpose_overlay_img/00000013.png
deleted file mode 100644
index 4e753362c07ffc7d5fb504aca0147096d9bb5412..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000013.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000014.png b/assets/WLASL_01214/dwpose_overlay_img/00000014.png
deleted file mode 100644
index a6dc064c213a8ac44855df4cbb27694658ba1ddd..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000014.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000015.png b/assets/WLASL_01214/dwpose_overlay_img/00000015.png
deleted file mode 100644
index d702fb4c190db35f3018ba8580314e872f46ca18..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000015.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000016.png b/assets/WLASL_01214/dwpose_overlay_img/00000016.png
deleted file mode 100644
index e79a7dd18b498cda02c4ed366a9cfb4c751e0487..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000016.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000017.png b/assets/WLASL_01214/dwpose_overlay_img/00000017.png
deleted file mode 100644
index c0bf8901eb696f906dfbf41c831296976602ffde..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000017.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000018.png b/assets/WLASL_01214/dwpose_overlay_img/00000018.png
deleted file mode 100644
index 07822b403763ce76f10625d607e9131b8580847d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000018.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000019.png b/assets/WLASL_01214/dwpose_overlay_img/00000019.png
deleted file mode 100644
index fb7f2625d3fc1c4ebcaab8a72d80c88daeaa898a..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000019.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000020.png b/assets/WLASL_01214/dwpose_overlay_img/00000020.png
deleted file mode 100644
index 14b93206ca8ca562b81123553286fbd70f2fa2c8..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000020.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000021.png b/assets/WLASL_01214/dwpose_overlay_img/00000021.png
deleted file mode 100644
index 47d1fe58f3df6a779efde374c542032daf0351fe..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000021.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000022.png b/assets/WLASL_01214/dwpose_overlay_img/00000022.png
deleted file mode 100644
index 38cc09e7ef5c068764764bbbdfdeee40a182e69e..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000022.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000023.png b/assets/WLASL_01214/dwpose_overlay_img/00000023.png
deleted file mode 100644
index a18c1cc69aad4fccfad644934433f92fbc824597..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000023.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000024.png b/assets/WLASL_01214/dwpose_overlay_img/00000024.png
deleted file mode 100644
index 79abf0063a549a42cfbc6356594ee364ba4631ed..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000024.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000025.png b/assets/WLASL_01214/dwpose_overlay_img/00000025.png
deleted file mode 100644
index 6ef84bf9ea8022f1af2c3e51ab5d50d401c1163a..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000025.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000026.png b/assets/WLASL_01214/dwpose_overlay_img/00000026.png
deleted file mode 100644
index 78e3c29f4c63f92e010a386c99450e94f22564e8..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000026.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000027.png b/assets/WLASL_01214/dwpose_overlay_img/00000027.png
deleted file mode 100644
index 9d479a05792fe42cd01f6c4c1faecb21a19ade38..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000027.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000028.png b/assets/WLASL_01214/dwpose_overlay_img/00000028.png
deleted file mode 100644
index 573c79b4804e5c1d94c83f3cf44c15dfb582316b..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000028.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000029.png b/assets/WLASL_01214/dwpose_overlay_img/00000029.png
deleted file mode 100644
index bdd6a25dae1b70b50e993819107e168e3244798e..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000029.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000030.png b/assets/WLASL_01214/dwpose_overlay_img/00000030.png
deleted file mode 100644
index 70eeecf9b7639d793aa30da6c2e48e98a0fe218c..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000030.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000031.png b/assets/WLASL_01214/dwpose_overlay_img/00000031.png
deleted file mode 100644
index 114810d3f296644296380f637d17bb4a07ac00fc..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000031.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000032.png b/assets/WLASL_01214/dwpose_overlay_img/00000032.png
deleted file mode 100644
index 3b7046ce11208f59cc8dbbba30730b780cca9e3d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000032.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000033.png b/assets/WLASL_01214/dwpose_overlay_img/00000033.png
deleted file mode 100644
index 6dd0bc19fa2219c1b18fe2226206b864052cd1e7..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000033.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000034.png b/assets/WLASL_01214/dwpose_overlay_img/00000034.png
deleted file mode 100644
index 79e93a36506af4ab0187d818ed0359294b57e8cb..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000034.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000035.png b/assets/WLASL_01214/dwpose_overlay_img/00000035.png
deleted file mode 100644
index a84a62d468d54c9492cbd264cae4dd34cc88a3fa..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000035.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000036.png b/assets/WLASL_01214/dwpose_overlay_img/00000036.png
deleted file mode 100644
index 0dc40405ee78d4c7b3dcbd9aca242689974df9bb..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000036.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000037.png b/assets/WLASL_01214/dwpose_overlay_img/00000037.png
deleted file mode 100644
index 4d1a938fbe8f193443f0cd581d642befce050498..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000037.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000038.png b/assets/WLASL_01214/dwpose_overlay_img/00000038.png
deleted file mode 100644
index 8e0b752175d729fcf96d182f1d2e939b23bb231c..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000038.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000039.png b/assets/WLASL_01214/dwpose_overlay_img/00000039.png
deleted file mode 100644
index 6b9e16048fa2781cf04d60e45dd4c92584ce991f..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000039.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000040.png b/assets/WLASL_01214/dwpose_overlay_img/00000040.png
deleted file mode 100644
index 904b47d332f924508589fb5fb414268a20b68983..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000040.png and /dev/null differ
diff --git a/assets/WLASL_01214/dwpose_overlay_img/00000041.png b/assets/WLASL_01214/dwpose_overlay_img/00000041.png
deleted file mode 100644
index 978ca46c5c6b238e4f3f3b1b01d0812707311990..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/dwpose_overlay_img/00000041.png and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000001.jpg b/assets/WLASL_01214/results_dwpose/img/00000001.jpg
deleted file mode 100644
index 2eba8dfba6bfe86474899ed1d15796ac7f0e5743..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000001.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000002.jpg b/assets/WLASL_01214/results_dwpose/img/00000002.jpg
deleted file mode 100644
index ea079b05ed26c0d9017770a7fd248c7fd04ce102..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000002.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000003.jpg b/assets/WLASL_01214/results_dwpose/img/00000003.jpg
deleted file mode 100644
index cf1f07341268fd2eec5fe10d6a34ef49e694c36d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000003.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000004.jpg b/assets/WLASL_01214/results_dwpose/img/00000004.jpg
deleted file mode 100644
index 4a1b7b5a86caa2176511bc4ac02d1de226b2480b..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000004.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000005.jpg b/assets/WLASL_01214/results_dwpose/img/00000005.jpg
deleted file mode 100644
index c6fb0582efc762cb0a30c8659551944097b836a1..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000005.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000006.jpg b/assets/WLASL_01214/results_dwpose/img/00000006.jpg
deleted file mode 100644
index d38c33a646544b093295b6a4e136b3d8fcda4ea9..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000006.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000007.jpg b/assets/WLASL_01214/results_dwpose/img/00000007.jpg
deleted file mode 100644
index 4d1633f14c50846a873e3a9087956d41e299ceb3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000007.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000008.jpg b/assets/WLASL_01214/results_dwpose/img/00000008.jpg
deleted file mode 100644
index c34bc8696d0dd61469787c1a0a1b99d31dda0148..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000008.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000009.jpg b/assets/WLASL_01214/results_dwpose/img/00000009.jpg
deleted file mode 100644
index eda34924dc830373b8b6ba5265444e2e2589a341..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000009.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000010.jpg b/assets/WLASL_01214/results_dwpose/img/00000010.jpg
deleted file mode 100644
index 1190d40abcd0b6cf25ab857cfa2a031f37e3ac2d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000010.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000011.jpg b/assets/WLASL_01214/results_dwpose/img/00000011.jpg
deleted file mode 100644
index f53eaa6b793bf4e8e62ad3cef403ad31679791c9..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000011.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000012.jpg b/assets/WLASL_01214/results_dwpose/img/00000012.jpg
deleted file mode 100644
index f2d86b6e91910f49e6cdd6f292455b70d030e029..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000012.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000013.jpg b/assets/WLASL_01214/results_dwpose/img/00000013.jpg
deleted file mode 100644
index d87a752a44a62ce455d4c3a12da97e259476f86e..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000013.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000014.jpg b/assets/WLASL_01214/results_dwpose/img/00000014.jpg
deleted file mode 100644
index eae09d1c107f5757e3e8a583162e55df2eeb5008..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000014.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000015.jpg b/assets/WLASL_01214/results_dwpose/img/00000015.jpg
deleted file mode 100644
index 443de07a00cf5edbb9846000fce82f6ab6908683..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000015.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000016.jpg b/assets/WLASL_01214/results_dwpose/img/00000016.jpg
deleted file mode 100644
index 26d9818c65027057b11df80d8f3ecae311549fe4..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000016.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000017.jpg b/assets/WLASL_01214/results_dwpose/img/00000017.jpg
deleted file mode 100644
index 7b7b27b92f2f900c816e11b10de61e90e9db5454..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000017.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000018.jpg b/assets/WLASL_01214/results_dwpose/img/00000018.jpg
deleted file mode 100644
index 115fb759c0ba0e9e5c94d55795f0e4c40f29e825..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000018.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000019.jpg b/assets/WLASL_01214/results_dwpose/img/00000019.jpg
deleted file mode 100644
index 18ea05721577b25b26cca90bd6311448a63cdf48..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000019.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000020.jpg b/assets/WLASL_01214/results_dwpose/img/00000020.jpg
deleted file mode 100644
index b02d018a7ba6a43895f7acdc083f4a95f2ff3acb..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000020.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000021.jpg b/assets/WLASL_01214/results_dwpose/img/00000021.jpg
deleted file mode 100644
index 9f7b67183f88ad24d1d21e37aa605bab3ce55115..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000021.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000022.jpg b/assets/WLASL_01214/results_dwpose/img/00000022.jpg
deleted file mode 100644
index f6007a3ed095e76d53a27b195fa21084734f6ea3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000022.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000023.jpg b/assets/WLASL_01214/results_dwpose/img/00000023.jpg
deleted file mode 100644
index bb105a06ce35e707bac734ed00ce3d3f2a5e0fee..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000023.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000024.jpg b/assets/WLASL_01214/results_dwpose/img/00000024.jpg
deleted file mode 100644
index db9ee561b7606ef64e546d29c2418fe09d790f25..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000024.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000025.jpg b/assets/WLASL_01214/results_dwpose/img/00000025.jpg
deleted file mode 100644
index 5e7b540669c5333dbb8dccb1dbf79586c7e1c0fd..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000025.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000026.jpg b/assets/WLASL_01214/results_dwpose/img/00000026.jpg
deleted file mode 100644
index e1eeb53db87f7c448196b821efb09967a340a01d..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000026.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000027.jpg b/assets/WLASL_01214/results_dwpose/img/00000027.jpg
deleted file mode 100644
index 0a77c289a79b218bbb160e5c5b4904e2e28fb401..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000027.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000028.jpg b/assets/WLASL_01214/results_dwpose/img/00000028.jpg
deleted file mode 100644
index 65d31d12c59ae351cb7888ee98e9e26935108039..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000028.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000029.jpg b/assets/WLASL_01214/results_dwpose/img/00000029.jpg
deleted file mode 100644
index 7078e4238a95e0b2bc376ca866dc0475b7a20fd3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000029.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000030.jpg b/assets/WLASL_01214/results_dwpose/img/00000030.jpg
deleted file mode 100644
index 9e2ec804561d6d1e990f582f82baa9b5ae287fe1..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000030.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000031.jpg b/assets/WLASL_01214/results_dwpose/img/00000031.jpg
deleted file mode 100644
index 6cc2333ad5392b4cb0752e658f5f02f8eb171d46..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000031.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000032.jpg b/assets/WLASL_01214/results_dwpose/img/00000032.jpg
deleted file mode 100644
index dacbf39595afb311c49eb2921562b24087c141a3..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000032.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000033.jpg b/assets/WLASL_01214/results_dwpose/img/00000033.jpg
deleted file mode 100644
index c25f9cd9105695df4af3b8ea13ab57e3ab956d31..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000033.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000034.jpg b/assets/WLASL_01214/results_dwpose/img/00000034.jpg
deleted file mode 100644
index 8907515df0efedce970402794d6c74183f185837..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000034.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000035.jpg b/assets/WLASL_01214/results_dwpose/img/00000035.jpg
deleted file mode 100644
index 2ff6c5e5fc2d4d72379d3236140d63c6590e4f88..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000035.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000036.jpg b/assets/WLASL_01214/results_dwpose/img/00000036.jpg
deleted file mode 100644
index f251f3a12a9e0c1d3364c01aed9518a82a016c17..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000036.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000037.jpg b/assets/WLASL_01214/results_dwpose/img/00000037.jpg
deleted file mode 100644
index fdbe96cb6a5e984ae02a72bcb54484f737ab6208..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000037.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000038.jpg b/assets/WLASL_01214/results_dwpose/img/00000038.jpg
deleted file mode 100644
index aa3466933651f397483a189effb208e4db3c91fc..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000038.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000039.jpg b/assets/WLASL_01214/results_dwpose/img/00000039.jpg
deleted file mode 100644
index 10505e9f775c00ba60f4ff72f52167cf023d7587..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000039.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000040.jpg b/assets/WLASL_01214/results_dwpose/img/00000040.jpg
deleted file mode 100644
index 821fc0db0d576fa40a1dc830efbd6c007de41b2c..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000040.jpg and /dev/null differ
diff --git a/assets/WLASL_01214/results_dwpose/img/00000041.jpg b/assets/WLASL_01214/results_dwpose/img/00000041.jpg
deleted file mode 100644
index 06ded3edb3d8aca6bf392ea59771b5f5fc99c493..0000000000000000000000000000000000000000
Binary files a/assets/WLASL_01214/results_dwpose/img/00000041.jpg and /dev/null differ
diff --git a/assets/pose.png b/assets/pose.png
deleted file mode 100644
index 63c4dcf8b564fe1f79d877a84787a7f1e19eaa5c..0000000000000000000000000000000000000000
Binary files a/assets/pose.png and /dev/null differ
diff --git a/assets/skeleton.png b/assets/skeleton.png
deleted file mode 100644
index be315d3e14944273b423d70761b4d8babed154a0..0000000000000000000000000000000000000000
Binary files a/assets/skeleton.png and /dev/null differ
diff --git a/easy_dwpose/README.md b/easy_dwpose/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..44632f513f93d08f01e65c62d9bc25fed7a697de
--- /dev/null
+++ b/easy_dwpose/README.md
@@ -0,0 +1,170 @@
+# fast_dwpose
+
+The goal of Easy DWPose is to provide a generic, reliable, and easy-to-use interface for making skeletons for ControlNet.
+
+SF do some improve for [easy-dwpose](https://github.com/reallyigor/easy_dwpose), named it fast-dwpose.
+
+## Installation
+
+### PIP
+
+```bash
+pip install easy-dwpose
+```
+
+## Quickstart
+
+### In you own .py scrip or in Jupyter
+
+```python
+import torch
+from PIL import Image
+import numpy as np
+import json
+
+from easy_dwpose import DWposeDetector
+
+#####---------Setup init
+# You can use a different GPU, e.g. "cuda:1"
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+detector = DWposeDetector(device=device)
+input_image = Image.open("assets/pose.png").convert("RGB")
+
+
+#####---------Get both the skeleton image
+# SF: skeleton should be a kind of img
+skeleton = detector(input_image, output_type="pil", include_hands=True, include_face=True)
+# Save the skeleton image
+skeleton.save("skeleton.png")
+
+
+#####---------Get pose data
+# SF: pose_data should be numpy/tensor
+# # This returns the dictionary
+pose_data = detector(input_image, draw_pose=False)
+
+# Save the skeleton pose information:
+# Option 1: Save as NPY file
+np.save('pose_data.npy', pose_data)
+
+# Option 2: Save as NPZ file
+np.savez('pose_data.npz', **pose_data)
+
+# Option 3: Save as JSON file
+# Convert numpy arrays to lists for JSON serialization
+pose_data_json = {k: v.tolist() if isinstance(v, np.ndarray) else v for k, v in pose_data.items()}
+with open('pose_data.json', 'w') as f:
+ json.dump(pose_data_json, f)
+```
+
+
+
+ | Input |
+ Output |
+
+
+
+
+
+ |
+
+
+
+ |
+
+
+
+### On a video
+
+```bash
+python scripts/inference_on_video.py --input assets/dance.mp4 --output_path result.mp4
+```
+
+
+
+ | Input |
+ Output |
+
+
+
+
+
+ |
+
+
+
+ |
+
+
+
+### On a folder of images
+
+```bash
+python scripts/inference_on_folder.py --input assets/ --output_path results/
+```
+
+### Easy-DWPose Custom skeleton drawing
+
+By default, we use standart skeleton drawing function but several projects change it (e.g. [MusePose](https://github.com/TMElyralab/MusePose)). Modify it or write your own from scratch!
+
+```python
+from PIL import Image
+from easy_dwpose import DWposeDetector
+from easy_dwpose.draw.musepose import draw_pose as draw_pose_musepose
+
+detector = DWposeDetector(device="cpu")
+input_image = Image.open("assets/pose.png").convert("RGB")
+
+skeleton = detector(input_image, output_type="pil", draw_pose=draw_pose_musepose, draw_face=False)
+skeleton.save("skeleton.png")
+```
+
+### SF Custom skeleton drawing
+
+I prefer ControlNext style, I have developed a visualization method and placed it in `./easy_dwpose/draw/controlnext.py`. It does not integrate with easy dwpose and the calling method is slightly different:
+
+```python
+import torch
+from PIL import Image
+import numpy as np
+import json
+
+from easy_dwpose import DWposeDetector
+from easy_dwpose.draw.controlnext import draw_pose, process_pose_data
+
+#####---------Setup init
+# You can use a different GPU, e.g. "cuda:1"
+device = "cuda:0" if torch.cuda.is_available() else "cpu"
+detector = DWposeDetector(device=device)
+input_image = Image.open("assets/pose.png").convert("RGB")
+
+#####---------Custom ControlNext drawing style
+# Get pose data for custom drawing
+pose_data = detector(input_image, draw_pose=False)
+
+# Get image dimensions
+width, height = input_image.size
+
+# Process the pose data for custom drawing
+processed_pred = process_pose_data(pose_data, height, width)
+
+# Draw pose using custom ControlNext style
+vis_img = draw_pose(
+ pose=processed_pred,
+ H=height,
+ W=width,
+ include_body=True,
+ include_hand=True,
+ include_face=True
+)
+
+# Convert to PIL Image and save (vis_img is in CHW format)
+custom_skeleton = Image.fromarray(vis_img.transpose(1, 2, 0))
+custom_skeleton.save("skeleton_controlnext.png")
+```
+
+## Acknowledgement
+
+We thank the original authors of the [DWPose](https://github.com/IDEA-Research/DWPose) for their incredible models!
+
+Thanks for open-sourcing!
diff --git a/env/environment01_dwpose.yml b/env/environment01_dwpose.yml
new file mode 100644
index 0000000000000000000000000000000000000000..358898fcf798d68b09ab2feb40bbf0b02e872ed6
--- /dev/null
+++ b/env/environment01_dwpose.yml
@@ -0,0 +1,13 @@
+name: dwpose
+channels:
+ - pytorch
+ - conda-forge
+ - nvidia
+ - defaults
+ - bioconda
+ - anaconda
+dependencies:
+ - python=3.10
+ - av
+ - numpy
+ - einops
\ No newline at end of file
diff --git a/env/requirements01_dwpose.txt b/env/requirements01_dwpose.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c0d244f71644620356177cfcdafbc1e43d675ed8
--- /dev/null
+++ b/env/requirements01_dwpose.txt
@@ -0,0 +1,5 @@
+easy_dwpose==1.0.1
+ffmpeg
+torch
+torchvision
+einops
\ No newline at end of file
diff --git a/step1_run_step1_from_csv.py b/pipeline/pipeline01_run_step1_from_csv.py
similarity index 100%
rename from step1_run_step1_from_csv.py
rename to pipeline/pipeline01_run_step1_from_csv.py
diff --git a/step2_merge_npz.py b/pipeline/pipeline02_merge_npz.py
similarity index 100%
rename from step2_merge_npz.py
rename to pipeline/pipeline02_merge_npz.py
diff --git a/step3_read_and_vis_merge_npz_to_pose_video.py b/pipeline/pipeline03_read_and_vis_merge_npz_to_pose_video.py
similarity index 97%
rename from step3_read_and_vis_merge_npz_to_pose_video.py
rename to pipeline/pipeline03_read_and_vis_merge_npz_to_pose_video.py
index 0e2e9c87288dffdfa2efbc3b07e47a8956e509f4..45ad853a6ef853078d1202851d403dc63c842e52 100644
--- a/step3_read_and_vis_merge_npz_to_pose_video.py
+++ b/pipeline/pipeline03_read_and_vis_merge_npz_to_pose_video.py
@@ -1,8 +1,8 @@
#!/usr/bin/env python3
"""
NPZ转Pose视频脚本
-用法: python step2_read_and_vis_merge_npz_to_pose_video.py
-例如: python step2_read_and_vis_merge_npz_to_pose_video.py ./posedict/WLASL_01214.npz
+用法: python pipeline/pipeline03_read_and_vis_merge_npz_to_pose_video.py
+例如: python pipeline/pipeline03_read_and_vis_merge_npz_to_pose_video.py ./posedict/WLASL_01214.npz
将合并后的多帧npz文件可视化成pose视频
"""
diff --git a/output/merge/database_summary.txt b/pose_dict/merge/database_summary.txt
similarity index 100%
rename from output/merge/database_summary.txt
rename to pose_dict/merge/database_summary.txt
diff --git a/output/merge_errors.txt b/pose_dict/merge_errors.txt
similarity index 100%
rename from output/merge_errors.txt
rename to pose_dict/merge_errors.txt
diff --git a/quick_vis_controlnext_style_dwpose_from_npz_or_input_img.py b/quick_vis_controlnext_style_dwpose_from_npz_or_input_img.py
deleted file mode 100644
index 3b3d95f5c7e77ef42800b260e6b6d24ead973146..0000000000000000000000000000000000000000
--- a/quick_vis_controlnext_style_dwpose_from_npz_or_input_img.py
+++ /dev/null
@@ -1,43 +0,0 @@
-import torch
-from PIL import Image
-import numpy as np
-import json
-
-from easy_dwpose import DWposeDetector
-from easy_dwpose.draw.controlnext import draw_pose, process_pose_data
-
-#####---------Setup init
-# You can use a different GPU, e.g. "cuda:1"
-device = "cuda:0" if torch.cuda.is_available() else "cpu"
-detector = DWposeDetector(device=device)
-input_image = Image.open("assets/pose.png").convert("RGB")
-
-
-
-#####---------Custom ControlNext drawing style
-# Get pose data for custom drawing
-#pose_data = detector(input_image, draw_pose=False)
-# or
-pose_data = dict(np.load('assets/WLASL_01214/results_dwpose/npz/00000011.npz'))
-
-# Get image dimensions
-#width, height = input_image.size
-# or
-width, height = 480, 480 #When I handle it, I default to a square with the largest side length. The size of the WLASL is 480*480.
-
-# Process the pose data for custom drawing
-processed_pred = process_pose_data(pose_data, height, width)
-
-# Draw pose using custom ControlNext style
-vis_img = draw_pose(
- pose=processed_pred,
- H=height,
- W=width,
- include_body=True,
- include_hand=True,
- include_face=True
-)
-
-# Convert to PIL Image and save (vis_img is in CHW format)
-custom_skeleton = Image.fromarray(vis_img.transpose(1, 2, 0))
-custom_skeleton.save("skeleton_controlnext.png")
\ No newline at end of file
diff --git a/skeleton_controlnext.png b/skeleton_controlnext.png
deleted file mode 100644
index 324aaa9447f93e5fe015f55b8d0c578228c48451..0000000000000000000000000000000000000000
Binary files a/skeleton_controlnext.png and /dev/null differ
diff --git a/tools-new-2025/README.md b/tools-new-2025/README.md
new file mode 100644
index 0000000000000000000000000000000000000000..5326e8656605c33e3de89fc62b282cbda9b38fdb
--- /dev/null
+++ b/tools-new-2025/README.md
@@ -0,0 +1,25 @@
+# Usage
+
+cd `Prompt2Sign/tools-new-2025/main`
+
+## Env
+
+```code
+conda env create -f env/environment01_dwpose.yml
+conda activate dwpose
+pip install -r env/requirements01_dwpose.txt
+pip install matplotlib-inline ipython
+conda install -n dwpose -c conda-forge ffmpeg -y
+```
+## Extract
+
+```code
+python pipeline00_extract_split_video_to_image.py
+python pipeline01_extract_dwpose_from_video.py
+
+# This requires a separate posture folder for each video, and obtaining the corresponding text description. Then, each line should correspond accordingly.
+# Since each dataset is different, please handle it by yourself.
+python pipeline02_obtain_trg384_data.py
+
+#Later, we realized that fully extending the posture and performing a sequence-to-sequence transformation might not be the optimal solution. Perhaps the third step was not very necessary.
+```
diff --git a/tools-new-2025/env/environment01_dwpose.yml b/tools-new-2025/env/environment01_dwpose.yml
new file mode 100644
index 0000000000000000000000000000000000000000..358898fcf798d68b09ab2feb40bbf0b02e872ed6
--- /dev/null
+++ b/tools-new-2025/env/environment01_dwpose.yml
@@ -0,0 +1,13 @@
+name: dwpose
+channels:
+ - pytorch
+ - conda-forge
+ - nvidia
+ - defaults
+ - bioconda
+ - anaconda
+dependencies:
+ - python=3.10
+ - av
+ - numpy
+ - einops
\ No newline at end of file
diff --git a/tools-new-2025/env/requirements01_dwpose.txt b/tools-new-2025/env/requirements01_dwpose.txt
new file mode 100644
index 0000000000000000000000000000000000000000..c0d244f71644620356177cfcdafbc1e43d675ed8
--- /dev/null
+++ b/tools-new-2025/env/requirements01_dwpose.txt
@@ -0,0 +1,5 @@
+easy_dwpose==1.0.1
+ffmpeg
+torch
+torchvision
+einops
\ No newline at end of file
diff --git a/tools-new-2025/input/WLASL_train/00335.mp4 b/tools-new-2025/input/WLASL_train/00335.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..e150b4acc07105a533f5b1c305cc8e2c731251d7
--- /dev/null
+++ b/tools-new-2025/input/WLASL_train/00335.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:36d2df8f69d774d0c1de525b4135b99b811a3fc0ebb7e321d4d0cb6facdc4ad1
+size 82741
diff --git a/tools-new-2025/input/WLASL_train/00339.mp4 b/tools-new-2025/input/WLASL_train/00339.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..f007aa819b23b3603b62308fbdb13ba9657a1be7
--- /dev/null
+++ b/tools-new-2025/input/WLASL_train/00339.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:1a8ad56112fb4d0f2f8e464d6689ca1ffd4793e0a51b8aec65020add714c8c34
+size 60415
diff --git a/tools-new-2025/input/WLASL_train/00376.mp4 b/tools-new-2025/input/WLASL_train/00376.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..ad65cf1abe6bf3c866a7fbcd4b631d8277da4a19
--- /dev/null
+++ b/tools-new-2025/input/WLASL_train/00376.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:83341144c0400236f92f0d63683f97df349969ce6a991da8c1236e6f70a87092
+size 81238
diff --git a/tools-new-2025/input/WLASL_train/00381.mp4 b/tools-new-2025/input/WLASL_train/00381.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..5c78cbe85342434e62eb5ca19bba6ba4e9844580
--- /dev/null
+++ b/tools-new-2025/input/WLASL_train/00381.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:47b29cf5c7db4198b4ee11fefc5f5b0e22d7aa82f4e9ab3b3239b19aeb26e487
+size 36967
diff --git a/tools-new-2025/input/test_dataset/human_mutiperson.mp4 b/tools-new-2025/input/test_dataset/human_mutiperson.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..615f14dd04fc2afa7ad25db2ceeb79b0af0fb788
--- /dev/null
+++ b/tools-new-2025/input/test_dataset/human_mutiperson.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:40071f3356f02d0c68624b7c331374dab4d23bf123cbeffd33906b58c85b5ff0
+size 16630806
diff --git a/tools-new-2025/input/test_dataset/human_signer.mp4 b/tools-new-2025/input/test_dataset/human_signer.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..64265aec60136d5a0ada3fe8a5128c787cd3aeaf
--- /dev/null
+++ b/tools-new-2025/input/test_dataset/human_signer.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:f9f19939db2854db63d6bb794ae25dce966ec290513ac39b4bbd5536cd7e4f20
+size 943389
diff --git a/tools-new-2025/input/test_dataset/human_signer_480.mp4 b/tools-new-2025/input/test_dataset/human_signer_480.mp4
new file mode 100644
index 0000000000000000000000000000000000000000..1bd636e758a3e1ad5bcec540d9fb1934f1da6476
--- /dev/null
+++ b/tools-new-2025/input/test_dataset/human_signer_480.mp4
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:d1df0a0dc33164ad1274beee742908d130bbc45840d916953c2294947c033fba
+size 415479
diff --git a/tools-new-2025/main/checkpoints/dw-ll_ucoco_384.onnx b/tools-new-2025/main/checkpoints/dw-ll_ucoco_384.onnx
new file mode 100644
index 0000000000000000000000000000000000000000..df84ce34881c5701a29e09badd8c96f5c17bd214
--- /dev/null
+++ b/tools-new-2025/main/checkpoints/dw-ll_ucoco_384.onnx
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:724f4ff2439ed61afb86fb8a1951ec39c6220682803b4a8bd4f598cd913b1843
+size 134399116
diff --git a/tools-new-2025/main/checkpoints/yolox_l.onnx b/tools-new-2025/main/checkpoints/yolox_l.onnx
new file mode 100644
index 0000000000000000000000000000000000000000..d6ff7914feb199e342967b877f8b2ea3179db915
--- /dev/null
+++ b/tools-new-2025/main/checkpoints/yolox_l.onnx
@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:7860ae79de6c89a3c1eb72ae9a2756c0ccfbe04b7791bb5880afabd97855a411
+size 216746733
diff --git a/tools-new-2025/main/pipeline00_extract_split_video_to_image.py b/tools-new-2025/main/pipeline00_extract_split_video_to_image.py
new file mode 100644
index 0000000000000000000000000000000000000000..ae41857c19c215da322710d4ea6f13901cb725d2
--- /dev/null
+++ b/tools-new-2025/main/pipeline00_extract_split_video_to_image.py
@@ -0,0 +1,187 @@
+import os
+import sys
+import subprocess
+from utils.preprocess_video import preprocess_videos, ensure_dir
+
+
+def main(dataset_name="test_dataset", square_crop=False, target_fps=1, quality_preset="medium", target_resolution=None):
+ """
+ Enhanced video preprocessing script - supports quality and resolution control
+
+ Args:
+ dataset_name (str): Dataset name
+ square_crop (bool): Whether to perform square cropping
+ target_fps (int): Target frame rate
+ quality_preset (str): Quality preset - "high", "medium", "low", "ultra_low"
+ target_resolution (int): Target resolution (pixels), None to keep original resolution
+ """
+ # Configure dataset name options
+ dataset_name = "test_dataset"
+ # dataset_name = "ASLLRP"
+ # dataset_name = "how2sign_test"
+ # dataset_name = "ncslgr_mp4"
+ # dataset_name = "ncslgr_mp4_small"
+ # dataset_name = "ASLLRP_batch_full_video_v1_1_small"
+ # dataset_name = "ASLLRP_utterances_small"
+ #dataset_name = "OpenVidHD-camera" # New large dataset
+
+ # === Configure processing parameters ===
+ square_crop = True # Whether to perform square cropping, False to keep original aspect ratio
+ target_fps = 30 # Target frame rate (recommended: 1-30)
+ quality_preset = "medium" # Quality preset: "high", "medium", "low", "ultra_low"
+ target_resolution = None # Target resolution (pixels), e.g.: 512, 256, None for original resolution
+
+ # === Common configuration combination examples ===
+
+ # 1. High quality processing (suitable for small datasets)
+ # quality_preset = "high"
+ # target_resolution = None # Keep original resolution
+
+ # 2. Balance quality and space (recommended)
+ # quality_preset = "medium"
+ # target_resolution = 512
+
+ # 3. Save space (suitable for large datasets)
+ # quality_preset = "low"
+ # target_resolution = 384
+
+ # 3.5 Maximum compression (suitable for very large datasets or limited storage)
+ #quality_preset = "low"
+ #target_resolution = 256
+
+ # 4. Maximum compression (suitable for very large datasets or limited storage)
+ # quality_preset = "ultra_low"
+ # target_resolution = 256
+
+ # 5. Special purpose: extremely low frame rate + small resolution (for rapid prototype testing)
+ # target_fps = 1
+ # quality_preset = "low"
+ # target_resolution = 224
+
+ # Set input and output directories
+ video_dir = ensure_dir(f"../input/{dataset_name}")
+ result_dir = ensure_dir(f"../output/{dataset_name}_results")
+
+ # Display configuration information
+ crop_status = "Square cropping" if square_crop else "Keep original aspect ratio"
+ resolution_info = f"{target_resolution}px" if target_resolution else "Original resolution"
+
+ print("=" * 60)
+ print("Enhanced Video Preprocessing")
+ print("=" * 60)
+ print(f"Dataset name: {dataset_name}")
+ print(f"Input directory: {video_dir}")
+ print(f"Output directory: {result_dir}")
+ print(f"Processing mode: {crop_status}")
+ print(f"Target frame rate: {target_fps} FPS")
+ print(f"Quality preset: {quality_preset}")
+ print(f"Target resolution: {resolution_info}")
+
+ # Estimate compression effect
+ compression_estimates = {
+ ("high", None): "File size approximately 80-100% of original",
+ ("medium", None): "File size approximately 40-60% of original",
+ ("low", None): "File size approximately 20-40% of original",
+ ("ultra_low", None): "File size approximately 10-20% of original",
+ ("medium", 512): "File size approximately 15-25% of original",
+ ("low", 384): "File size approximately 8-15% of original",
+ ("ultra_low", 256): "File size approximately 3-8% of original"
+ }
+
+ estimate_key = (quality_preset, target_resolution)
+ if estimate_key in compression_estimates:
+ print(f"Estimated effect: {compression_estimates[estimate_key]}")
+ else:
+ print("Estimated effect: Custom compression based on settings")
+
+ print("=" * 60)
+
+ # Execute video preprocessing
+ try:
+ preprocess_videos(
+ video_dir=video_dir,
+ dataset_name=dataset_name,
+ square_crop=square_crop,
+ fps=target_fps,
+ quality_preset=quality_preset,
+ target_resolution=target_resolution
+ )
+ print("\n" + "=" * 60)
+ print("Video preprocessing completed!")
+
+ # Statistics of processing results
+ processed_count = 0
+ total_size = 0
+
+ for video_file in os.listdir(video_dir):
+ if video_file.endswith(".mp4"):
+ video_name = os.path.splitext(video_file)[0]
+ output_video_path = os.path.join(result_dir, video_name, "crop_original_video.mp4")
+
+ if os.path.exists(output_video_path):
+ processed_count += 1
+ total_size += os.path.getsize(output_video_path)
+
+ total_size_gb = total_size / (1024**3)
+
+ print(f"Successfully processed {processed_count} videos")
+ print(f"Total output video size: {total_size_gb:.2f} GB")
+ print("=" * 60)
+
+ except Exception as e:
+ print(f"Error occurred during preprocessing: {str(e)}")
+
+
+def print_usage():
+ """Print usage instructions"""
+ print("""
+Usage:
+1. Basic usage (with default parameters):
+ python script.py
+
+2. Specify dataset:
+ python script.py dataset_name
+
+3. Full parameters:
+ python script.py dataset_name square_crop fps quality_preset resolution
+
+Parameter descriptions:
+- dataset_name: Dataset name (e.g.: "test_dataset")
+- square_crop: true/false (whether to perform square cropping)
+- fps: Frame rate (e.g.: 30, 15, 1)
+- quality_preset: high/medium/low/ultra_low (quality preset)
+- resolution: Target resolution in pixels (e.g.: 512, 256, or none to keep original resolution)
+
+Examples:
+python script.py OpenVidHD false 1 low 384
+python script.py test_dataset true 30 medium 512
+python script.py my_dataset false 15 ultra_low 256
+
+Recommended configurations:
+- Small datasets: medium quality + original resolution
+- Large datasets: low quality + 512 resolution
+- Very large datasets: ultra_low quality + 256 resolution
+""")
+
+
+if __name__ == '__main__':
+ if len(sys.argv) > 1:
+ if sys.argv[1] in ['-h', '--help', 'help']:
+ print_usage()
+ sys.exit(0)
+
+ dataset_name = sys.argv[1]
+ square_crop = sys.argv[2].lower() == 'true' if len(sys.argv) > 2 else False
+ target_fps = int(sys.argv[3]) if len(sys.argv) > 3 else 1
+ quality_preset = sys.argv[4] if len(sys.argv) > 4 else "medium"
+
+ # Handle resolution parameter
+ if len(sys.argv) > 5:
+ resolution_str = sys.argv[5].lower()
+ target_resolution = None if resolution_str in ['none', 'null', 'original'] else int(resolution_str)
+ else:
+ target_resolution = None
+
+ main(dataset_name, square_crop, target_fps, quality_preset, target_resolution)
+ else:
+ main()
\ No newline at end of file
diff --git a/tools-new-2025/main/pipeline01_extract_dwpose_from_video.py b/tools-new-2025/main/pipeline01_extract_dwpose_from_video.py
new file mode 100644
index 0000000000000000000000000000000000000000..bb832d5f02e7e48e138b9b677d3abcb6d401469a
--- /dev/null
+++ b/tools-new-2025/main/pipeline01_extract_dwpose_from_video.py
@@ -0,0 +1,316 @@
+import os
+import subprocess
+import torch
+import torch.distributed as dist
+import torch.multiprocessing as mp
+import numpy as np
+import json
+from easy_dwpose import DWposeDetector
+from PIL import Image
+from utils.util import read_frames
+from utils.draw_dw_lib import draw_pose, process_pose_data
+
+# Set distributed training environment variables
+os.environ["NCCL_DEBUG"] = "INFO"
+os.environ["NCCL_IB_DISABLE"] = "1"
+os.environ["NCCL_P2P_DISABLE"] = "1"
+
+def setup(rank, world_size):
+ """Setup distributed training environment"""
+ try:
+ os.environ['MASTER_ADDR'] = '127.0.0.1'
+ os.environ['MASTER_PORT'] = '23457' # Use different port to avoid conflicts
+ dist.init_process_group(
+ "gloo",
+ rank=rank,
+ world_size=world_size
+ )
+ return True
+ except Exception as e:
+ print(f"Failed to initialize process group: {str(e)}")
+ return False
+
+def cleanup():
+ """Clean up distributed training environment"""
+ dist.destroy_process_group()
+
+def save_dwpose_data(pose_data, frame_idx, json_dir, npy_dir, npz_dir, img_dir, frame_pil, detector):
+ """Save DWpose data to various formats - completely use new data structure"""
+ try:
+ frame_name = f"{frame_idx:08d}"
+
+ # Detect number of persons
+ num_persons = pose_data['faces'].shape[0] if 'faces' in pose_data else 0
+
+ # Completely use new data structure
+ new_pose_data = {
+ 'num_persons': num_persons
+ }
+
+ if num_persons > 0:
+ # Reorganize body keypoint data
+ bodies_reshaped = pose_data['bodies'].reshape(num_persons, 18, 2)
+
+ # Add separate data for each person
+ for person_id in range(num_persons):
+ new_pose_data[f'person_{person_id}'] = {
+ 'body_keypoints': bodies_reshaped[person_id],
+ 'body_scores': pose_data['body_scores'][person_id],
+ 'face_keypoints': pose_data['faces'][person_id],
+ 'face_scores': pose_data['faces_scores'][person_id],
+ 'left_hand_keypoints': pose_data['hands'][person_id * 2],
+ 'left_hand_scores': pose_data['hands_scores'][person_id * 2],
+ 'right_hand_keypoints': pose_data['hands'][person_id * 2 + 1],
+ 'right_hand_scores': pose_data['hands_scores'][person_id * 2 + 1],
+ }
+
+ # Save NPY file - only save new format
+ try:
+ np.save(os.path.join(npy_dir, f"{frame_name}.npy"), new_pose_data)
+ except Exception as e:
+ print(f"Failed to save NPY file (frame {frame_idx}): {str(e)}")
+
+ # Save NPZ file - only save new format
+ try:
+ np.savez(os.path.join(npz_dir, f"{frame_name}.npz"), **new_pose_data)
+ except Exception as e:
+ print(f"Failed to save NPZ file (frame {frame_idx}): {str(e)}")
+
+ # Save visualization image
+ try:
+ # Use new visualization method
+ try:
+ width, height = frame_pil.size
+ processed_pred = process_pose_data(new_pose_data, height, width)
+
+ vis_img = draw_pose(
+ pose=processed_pred,
+ H=height,
+ W=width,
+ include_body=True,
+ include_hand=True,
+ include_face=True
+ )
+
+ # Handle channel order
+ if isinstance(vis_img, np.ndarray):
+ if vis_img.shape[0] == 3 and len(vis_img.shape) == 3:
+ vis_img = np.transpose(vis_img, (1, 2, 0))
+
+ if vis_img.dtype != np.uint8:
+ vis_img = (vis_img * 255).astype(np.uint8)
+
+ pred_img = Image.fromarray(vis_img)
+ else:
+ raise ValueError(f"draw_pose returned invalid type: {type(vis_img)}")
+
+ except Exception as viz_error:
+ print(f"Warning: New visualization method failed ({str(viz_error)}), using original method")
+ pred_img = detector(frame_pil, output_type="pil", include_hands=True, include_face=True)
+
+ pred_img.save(os.path.join(img_dir, f"{frame_name}.jpg"))
+
+ except Exception as e:
+ print(f"Failed to save pose visualization image (frame {frame_idx}): {str(e)}")
+
+ # Create overlay visualization image
+ try:
+ overlay_dir = os.path.join(os.path.dirname(os.path.dirname(img_dir)), "dwpose_overlay_img")
+ os.makedirs(overlay_dir, exist_ok=True)
+
+ processed_pred = process_pose_data(new_pose_data, height, width)
+ transparent_vis = draw_pose(
+ pose=processed_pred,
+ H=height,
+ W=width,
+ include_body=True,
+ include_hand=True,
+ include_face=True,
+ transparent=True
+ )
+
+ if isinstance(transparent_vis, np.ndarray):
+ if transparent_vis.shape[0] == 4 and len(transparent_vis.shape) == 3:
+ transparent_vis = np.transpose(transparent_vis, (1, 2, 0))
+ transparent_img = Image.fromarray(transparent_vis, 'RGBA')
+
+ frame_rgba = frame_pil.convert('RGBA')
+ overlay_img = Image.alpha_composite(frame_rgba, transparent_img)
+
+ overlay_path = os.path.join(overlay_dir, f"{frame_name}.png")
+ overlay_img.save(overlay_path)
+
+ except Exception as e:
+ print(f"Failed to create overlay visualization (frame {frame_idx}): {str(e)}")
+
+ # Save JSON file - only save new format
+ try:
+ pose_data_json = {}
+ for k, v in new_pose_data.items():
+ if isinstance(v, np.ndarray):
+ pose_data_json[k] = v.tolist()
+ elif isinstance(v, (np.int32, np.int64)):
+ pose_data_json[k] = int(v)
+ elif isinstance(v, (np.float32, np.float64)):
+ pose_data_json[k] = float(v)
+ elif isinstance(v, dict):
+ # Handle person_id dictionary
+ person_dict = {}
+ for pk, pv in v.items():
+ if isinstance(pv, np.ndarray):
+ person_dict[pk] = pv.tolist()
+ else:
+ person_dict[pk] = pv
+ pose_data_json[k] = person_dict
+ else:
+ pose_data_json[k] = v
+
+ json_path = os.path.join(json_dir, f"{frame_name}.json")
+ with open(json_path, 'w') as f:
+ json.dump(pose_data_json, f)
+
+ except Exception as e:
+ print(f"Failed to save JSON file (frame {frame_idx}): {str(e)}")
+
+ except Exception as e:
+ print(f"General error in save_dwpose_data (frame {frame_idx}): {str(e)}")
+
+def process_video_with_dwpose(rank, video_path, output_base_dir, fps):
+ """Process single video using specified GPU"""
+ json_dir = os.path.join(output_base_dir, "json")
+ npy_dir = os.path.join(output_base_dir, "npy")
+ npz_dir = os.path.join(output_base_dir, "npz")
+ img_dir = os.path.join(output_base_dir, "img")
+
+ os.makedirs(json_dir, exist_ok=True)
+ os.makedirs(npy_dir, exist_ok=True)
+ os.makedirs(npz_dir, exist_ok=True)
+ os.makedirs(img_dir, exist_ok=True)
+
+ print(f"GPU {rank}: DWpose - Output directories created")
+
+ device = f"cuda:{rank}"
+ detector = DWposeDetector(device=device)
+ print(f"GPU {rank}: DWpose - Detector initialized on {device}")
+
+ frames = read_frames(video_path)
+ total_frames = len(frames)
+ print(f"GPU {rank}: DWpose - Read {total_frames} frames from video")
+
+ for i, frame_pil in enumerate(frames):
+ pose_data = detector(frame_pil, draw_pose=False, include_hands=True, include_face=True)
+ save_dwpose_data(pose_data, i + 1, json_dir, npy_dir, npz_dir, img_dir, frame_pil, detector)
+
+ if (i + 1) % 50 == 0 or i == 0 or i == total_frames - 1:
+ print(f"GPU {rank}: DWpose - Processed {i + 1}/{total_frames} frames ({(i + 1)/total_frames*100:.1f}%)")
+
+ # Generate visualization video
+ visualization_path = os.path.join(os.path.dirname(output_base_dir), "visualization_dwpose.mp4")
+ print(f"GPU {rank}: DWpose - Creating visualization video...")
+ try:
+ cmd = f'ffmpeg -y -framerate {fps} -i {img_dir}/%08d.jpg -c:v libx264 -pix_fmt yuv420p {visualization_path}'
+ subprocess.run(cmd, shell=True, check=True)
+ print(f"GPU {rank}: DWpose - Successfully created visualization video: {visualization_path}")
+ except Exception as e:
+ print(f"GPU {rank}: DWpose - Failed to create visualization video: {str(e)}")
+
+ # Generate overlay visualization video
+ overlay_dir = os.path.join(os.path.dirname(output_base_dir), "dwpose_overlay_img")
+ visualization_overlap_path = os.path.join(os.path.dirname(output_base_dir), "visualization_dwpose_overlap.mp4")
+ print(f"GPU {rank}: DWpose - Creating overlay visualization video...")
+ try:
+ cmd = f'ffmpeg -y -framerate {fps} -i {overlay_dir}/%08d.png -c:v libx264 -pix_fmt yuv420p {visualization_overlap_path}'
+ subprocess.run(cmd, shell=True, check=True)
+ print(f"GPU {rank}: DWpose - Successfully created overlay visualization video: {visualization_overlap_path}")
+ except Exception as e:
+ print(f"GPU {rank}: DWpose - Failed to create overlay visualization video: {str(e)}")
+
+def process_videos_distributed(rank, world_size, video_paths, dataset_name, fps):
+ """Process videos in distributed manner"""
+ if not setup(rank, world_size):
+ return
+
+ result_dir = f"../output/{dataset_name}_results"
+
+ for i, (video_name, video_path) in enumerate(video_paths):
+ if i % world_size == rank:
+ print(f"GPU {rank} processing video: {video_name}")
+
+ try:
+ folder_path = f"{result_dir}/{video_name}"
+ dwpose_results_dir = f"{folder_path}/results_dwpose"
+
+ if not os.path.exists(os.path.join(dwpose_results_dir, "json")):
+ os.makedirs(dwpose_results_dir, exist_ok=True)
+ print(f"GPU {rank}: Created directory {dwpose_results_dir}")
+ process_video_with_dwpose(rank, video_path, dwpose_results_dir, fps)
+ print(f"GPU {rank}: Completed DWpose processing for {video_name}")
+
+ with open(f"{result_dir}/processed_video_name_by_dwpose.txt", "a") as f:
+ f.write(f"{video_name}\n")
+ else:
+ print(f"GPU {rank}: {video_name} already processed, skipping")
+
+ except Exception as e:
+ print(f"GPU {rank}: Error processing video {video_name}: {str(e)}")
+ continue
+
+ cleanup()
+
+def main():
+ """Main function"""
+ # Configure dataset name
+ dataset_name = "test_dataset"
+ # dataset_name = "how2sign_test"
+ # dataset_name = "ncslgr_mp4"
+ # dataset_name = "ncslgr_mp4_small"
+ # dataset_name = "ASLLRP_batch_full_video_v1_1_small"
+ # dataset_name = "ASLLRP_utterances_small"
+
+ result_dir = f"../output/{dataset_name}_results"
+
+ fps = 24
+
+ # Collect videos to be processed
+ video_paths = []
+ processed_dwpose = set()
+
+ # Read list of already processed videos
+ dwpose_log = f"{result_dir}/processed_video_name_by_dwpose.txt"
+
+ if os.path.exists(dwpose_log):
+ with open(dwpose_log, 'r') as f:
+ processed_dwpose = set(f.read().splitlines())
+
+ # Check videos that need DWpose processing
+ video_dir = f"../input/{dataset_name}"
+ for video_file in os.listdir(video_dir):
+ if video_file.endswith(".mp4"):
+ video_name = os.path.splitext(video_file)[0]
+ crop_video_path = os.path.join(result_dir, video_name, "crop_original_video.mp4")
+
+ if os.path.exists(crop_video_path) and video_name not in processed_dwpose:
+ video_paths.append((video_name, crop_video_path))
+ print(f"Added video for DWpose processing: {video_name}")
+
+ # Run DWpose processing
+ if video_paths:
+ print(f"Found {len(video_paths)} videos that need DWpose processing")
+
+ world_size = torch.cuda.device_count()
+ if world_size > 1:
+ print(f"Found {world_size} GPUs, running distributed processing")
+ mp.spawn(
+ process_videos_distributed,
+ args=(world_size, video_paths, dataset_name, fps),
+ nprocs=world_size,
+ join=True
+ )
+ else:
+ print("Only one GPU available, running single GPU mode")
+ process_videos_distributed(0, 1, video_paths, dataset_name, fps)
+ else:
+ print("All videos have already been processed with DWpose")
+
+if __name__ == '__main__':
+ main()
\ No newline at end of file
diff --git a/tools-new-2025/main/utils/__pycache__/draw_dw_lib.cpython-310.pyc b/tools-new-2025/main/utils/__pycache__/draw_dw_lib.cpython-310.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..b6c2c6f6b9f41e393550aab6aca01056d3395c5e
Binary files /dev/null and b/tools-new-2025/main/utils/__pycache__/draw_dw_lib.cpython-310.pyc differ
diff --git a/tools-new-2025/main/utils/__pycache__/preprocess_video.cpython-310.pyc b/tools-new-2025/main/utils/__pycache__/preprocess_video.cpython-310.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..dbbdd400b9dc3fb101fbe554d42d564fe66c362d
Binary files /dev/null and b/tools-new-2025/main/utils/__pycache__/preprocess_video.cpython-310.pyc differ
diff --git a/tools-new-2025/main/utils/__pycache__/util.cpython-310.pyc b/tools-new-2025/main/utils/__pycache__/util.cpython-310.pyc
new file mode 100644
index 0000000000000000000000000000000000000000..67a0cddffbf929daad8145be0e8cdee6864fbec9
Binary files /dev/null and b/tools-new-2025/main/utils/__pycache__/util.cpython-310.pyc differ
diff --git a/tools-new-2025/main/utils/draw_dw_lib.py b/tools-new-2025/main/utils/draw_dw_lib.py
new file mode 100644
index 0000000000000000000000000000000000000000..d75b9aa4fcf79b07afc972482a61028d4524cd26
--- /dev/null
+++ b/tools-new-2025/main/utils/draw_dw_lib.py
@@ -0,0 +1,294 @@
+import math
+import numpy as np
+import matplotlib
+import cv2
+import sys
+import os
+import _pickle as cPickle
+import gzip
+import subprocess
+import torch
+import colorsys
+from typing import List, Dict, Any, Optional, Tuple
+
+
+eps = 0.01
+
+def alpha_blend_color(color, alpha):
+ """blend color according to point conf
+ """
+ return [int(c * alpha) for c in color]
+
+def draw_bodypose(canvas, candidate, subset, score, transparent=False):
+ """Draw body pose on canvas
+ Args:
+ canvas: numpy array canvas to draw on
+ candidate: pose candidate
+ subset: pose subset
+ score: confidence scores
+ transparent: whether to use transparent background
+ Returns:
+ canvas: drawn canvas
+ """
+ H, W, C = canvas.shape
+ candidate = np.array(candidate)
+ subset = np.array(subset)
+
+ stickwidth = 4
+
+ limbSeq = [[2, 3], [2, 6], [3, 4], [4, 5], [6, 7], [7, 8], [2, 9], [9, 10],
+ [10, 11], [2, 12], [12, 13], [13, 14], [2, 1], [1, 15], [15, 17],
+ [1, 16], [16, 18], [3, 17], [6, 18]]
+
+ colors = [[255, 0, 0], [255, 85, 0], [255, 170, 0], [255, 255, 0], [170, 255, 0], [85, 255, 0], [0, 255, 0],
+ [0, 255, 85], [0, 255, 170], [0, 255, 255], [0, 170, 255], [0, 85, 255], [0, 0, 255], [85, 0, 255],
+ [170, 0, 255], [255, 0, 255], [255, 0, 170], [255, 0, 85]]
+
+ # Add alpha channel if transparent
+ if transparent:
+ colors = [color + [255] for color in colors]
+
+ for i in range(17):
+ for n in range(len(subset)):
+ index = subset[n][np.array(limbSeq[i]) - 1]
+ conf = score[n][np.array(limbSeq[i]) - 1]
+ if conf[0] < 0.3 or conf[1] < 0.3:
+ continue
+ Y = candidate[index.astype(int), 0] * float(W)
+ X = candidate[index.astype(int), 1] * float(H)
+ mX = np.mean(X)
+ mY = np.mean(Y)
+ length = ((X[0] - X[1]) ** 2 + (Y[0] - Y[1]) ** 2) ** 0.5
+ angle = math.degrees(math.atan2(X[0] - X[1], Y[0] - Y[1]))
+ polygon = cv2.ellipse2Poly((int(mY), int(mX)), (int(length / 2), stickwidth), int(angle), 0, 360, 1)
+ if transparent:
+ color = colors[i][:-1] + [int(255 * conf[0] * conf[1])] # Adjust alpha based on confidence
+ else:
+ color = colors[i]
+ cv2.fillConvexPoly(canvas, polygon, color)
+
+ canvas = (canvas * 0.6).astype(np.uint8)
+
+ for i in range(18):
+ for n in range(len(subset)):
+ index = int(subset[n][i])
+ if index == -1:
+ continue
+ x, y = candidate[index][0:2]
+ conf = score[n][i]
+ x = int(x * W)
+ y = int(y * H)
+ if transparent:
+ color = colors[i][:-1] + [int(255 * conf)] # Adjust alpha based on confidence
+ else:
+ color = colors[i]
+ cv2.circle(canvas, (int(x), int(y)), 4, color, thickness=-1)
+
+ return canvas
+
+def draw_handpose(canvas, all_hand_peaks, all_hand_scores, transparent=False):
+ """Draw hand pose on canvas"""
+ H, W, C = canvas.shape
+
+ edges = [[0, 1], [1, 2], [2, 3], [3, 4], [0, 5], [5, 6], [6, 7], [7, 8], [0, 9], [9, 10],
+ [10, 11], [11, 12], [0, 13], [13, 14], [14, 15], [15, 16], [0, 17], [17, 18], [18, 19], [19, 20]]
+
+ for peaks, scores in zip(all_hand_peaks, all_hand_scores):
+ for ie, e in enumerate(edges):
+ x1, y1 = peaks[e[0]]
+ x2, y2 = peaks[e[1]]
+ x1 = int(x1 * W)
+ y1 = int(y1 * H)
+ x2 = int(x2 * W)
+ y2 = int(y2 * H)
+ score = scores[e[0]] * scores[e[1]]
+ if x1 > eps and y1 > eps and x2 > eps and y2 > eps:
+ color = matplotlib.colors.hsv_to_rgb([ie / float(len(edges)), 1.0, 1.0])
+ if transparent:
+ color = np.append(color, score) # Add alpha channel
+ else:
+ color = color * score
+ cv2.line(canvas, (x1, y1), (x2, y2), color * 255, thickness=2)
+
+ for i, keypoint in enumerate(peaks):
+ x, y = keypoint
+ x = int(x * W)
+ y = int(y * H)
+ if x > eps and y > eps:
+ if transparent:
+ color = (0, 0, 0, scores[i]) # Black with alpha
+ else:
+ color = (0, 0, int(scores[i] * 255)) # Original color
+ cv2.circle(canvas, (x, y), 4, color, thickness=-1)
+ return canvas
+
+def draw_facepose(canvas, all_lmks, all_scores, transparent=False):
+ """Draw face pose on canvas"""
+ H, W, C = canvas.shape
+ for lmks, scores in zip(all_lmks, all_scores):
+ for lmk, score in zip(lmks, scores):
+ x, y = lmk
+ x = int(x * W)
+ y = int(y * H)
+ if x > eps and y > eps:
+ if transparent:
+ color = (255, 255, 255, int(score * 255)) # White with alpha
+ else:
+ conf = int(score * 255)
+ color = (conf, conf, conf) # Original grayscale
+ cv2.circle(canvas, (x, y), 3, color, thickness=-1)
+ return canvas
+
+def draw_pose(pose, H, W, include_body=True, include_hand=True, include_face=True, ref_w=2160, transparent=False):
+ """vis dwpose outputs with optional transparent background
+
+ Args:
+ pose (Dict): DWposeDetector outputs - 支持新的person_id格式和旧格式
+ H (int): height
+ W (int): width
+ include_body (bool): whether to draw body keypoints
+ include_hand (bool): whether to draw hand keypoints
+ include_face (bool): whether to draw face keypoints
+ ref_w (int, optional): reference width. Defaults to 2160.
+ transparent (bool, optional): whether to use transparent background. Defaults to False.
+
+ Returns:
+ np.ndarray: image pixel value in RGBA mode if transparent=True, otherwise RGB mode
+ """
+ sz = min(H, W)
+ sr = (ref_w / sz) if sz != ref_w else 1
+
+ # Create canvas - now with alpha channel if transparent
+ if transparent:
+ canvas = np.zeros(shape=(int(H*sr), int(W*sr), 4), dtype=np.uint8)
+ else:
+ canvas = np.zeros(shape=(int(H*sr), int(W*sr), 3), dtype=np.uint8)
+
+ # 检查是否是新的person_id数据格式
+ if 'num_persons' in pose and pose['num_persons'] > 0:
+ # 使用新的多人数据结构
+ processed_data = process_pose_data(pose, H, W)
+ bodies = processed_data['bodies']
+ faces = processed_data['faces']
+ hands = processed_data['hands']
+ candidate = bodies['candidate']
+ subset = bodies['subset']
+
+ if include_body:
+ canvas = draw_bodypose(canvas, candidate, subset, score=bodies['score'], transparent=transparent)
+
+ if include_hand:
+ canvas = draw_handpose(canvas, hands, processed_data['hands_score'], transparent=transparent)
+
+ if include_face:
+ canvas = draw_facepose(canvas, faces, processed_data['faces_score'], transparent=transparent)
+
+ else:
+ # 兼容旧的数据格式 - 作为备选方案
+ try:
+ bodies = pose['bodies']
+ faces = pose['faces']
+ hands = pose['hands']
+ candidate = bodies['candidate']
+ subset = bodies['subset']
+
+ if include_body:
+ canvas = draw_bodypose(canvas, candidate, subset, score=bodies['score'], transparent=transparent)
+
+ if include_hand:
+ canvas = draw_handpose(canvas, hands, pose['hands_score'], transparent=transparent)
+
+ if include_face:
+ canvas = draw_facepose(canvas, faces, pose['faces_score'], transparent=transparent)
+ except Exception as e:
+ print(f"绘制旧格式数据失败: {str(e)}")
+ # 返回空画布
+ pass
+
+ if transparent:
+ return cv2.cvtColor(cv2.resize(canvas, (W, H)), cv2.COLOR_BGRA2RGBA).transpose(2, 0, 1)
+ else:
+ return cv2.cvtColor(cv2.resize(canvas, (W, H)), cv2.COLOR_BGR2RGB).transpose(2, 0, 1)
+
+def process_pose_data(pose_data: Dict[str, Any], height: int, width: int) -> Dict[str, Any]:
+ """
+ 处理姿势数据，完全支持新的person_id数据结构
+ """
+ processed_data = {}
+
+ # 确保使用新的数据结构
+ if 'num_persons' in pose_data and pose_data['num_persons'] > 0:
+ num_persons = pose_data['num_persons']
+
+ # 收集所有人的关键点数据
+ all_bodies = []
+ all_body_scores = []
+ all_hands = []
+ all_hand_scores = []
+ all_faces = []
+ all_face_scores = []
+
+ for person_id in range(num_persons):
+ person_key = f'person_{person_id}'
+ if person_key in pose_data:
+ person_data = pose_data[person_key]
+ all_bodies.append(person_data['body_keypoints'])
+ all_body_scores.append(person_data['body_scores'])
+ all_hands.extend([person_data['left_hand_keypoints'], person_data['right_hand_keypoints']])
+ all_hand_scores.extend([person_data['left_hand_scores'], person_data['right_hand_scores']])
+ all_faces.append(person_data['face_keypoints'])
+ all_face_scores.append(person_data['face_scores'])
+
+ # 合并所有人的数据
+ if all_bodies:
+ bodies = np.vstack(all_bodies)
+ body_scores = np.array(all_body_scores)
+
+ # 创建subset - 为每个人创建独立的subset行
+ subset = []
+ for person_id in range(num_persons):
+ person_subset = list(range(person_id * 18, (person_id + 1) * 18))
+ subset.append(person_subset)
+ subset = np.array(subset)
+
+ # 创建scores - 基于body_scores中的有效性
+ scores = np.ones_like(body_scores)
+ for i in range(num_persons):
+ for j in range(18):
+ if body_scores[i, j] < 0: # 如果body_scores为负数，认为无效
+ scores[i, j] = 0.0
+ else:
+ scores[i, j] = 1.0
+ else:
+ bodies = np.array([])
+ subset = np.array([[]])
+ scores = np.array([[]])
+
+ hands = np.array(all_hands) if all_hands else np.array([])
+ hand_scores = np.array(all_hand_scores) if all_hand_scores else np.array([])
+ faces = np.array(all_faces) if all_faces else np.array([])
+ face_scores = np.array(all_face_scores) if all_face_scores else np.array([])
+
+ else:
+ # 兼容性处理 - 如果不是新格式，返回空数据
+ bodies = np.array([])
+ subset = np.array([[]])
+ scores = np.array([[]])
+ hands = np.array([])
+ hand_scores = np.array([])
+ faces = np.array([])
+ face_scores = np.array([])
+
+ processed_data['bodies'] = {
+ 'candidate': bodies,
+ 'subset': subset,
+ 'score': scores
+ }
+
+ processed_data['hands'] = hands
+ processed_data['hands_score'] = hand_scores
+
+ processed_data['faces'] = faces
+ processed_data['faces_score'] = face_scores
+
+ return processed_data
\ No newline at end of file
diff --git a/tools-new-2025/main/utils/preprocess_video.py b/tools-new-2025/main/utils/preprocess_video.py
new file mode 100644
index 0000000000000000000000000000000000000000..611969a0ca9d39c2e907f5866388619cacd66f7a
--- /dev/null
+++ b/tools-new-2025/main/utils/preprocess_video.py
@@ -0,0 +1,259 @@
+import os
+import shutil
+import subprocess
+import av
+import torch
+import torch.distributed as dist
+import torch.multiprocessing as mp
+from torch.nn.parallel import DistributedDataParallel
+from utils.util import get_fps, read_frames, save_videos_from_pil
+from utils.preprocess_video import *
+from PIL import Image
+import numpy as np
+import json
+
+def ensure_dir(directory):
+ if os.path.exists(directory):
+ print(f"Directory already exists: {directory}")
+ else:
+ os.makedirs(directory)
+ print(f"Created directory: {directory}")
+ return directory
+
+# [previous helper functions remain the same]
+def get_video_dimensions(video_path):
+ cmd = [
+ 'ffprobe',
+ '-v', 'error',
+ '-select_streams', 'v:0',
+ '-show_entries', 'stream=width,height',
+ '-of', 'csv=p=0',
+ video_path
+ ]
+ result = subprocess.run(cmd, capture_output=True, text=True)
+ width, height = map(int, result.stdout.strip().split(','))
+ return width, height
+
+def compile_frames_to_video(frame_dir, output_path, fps=30):
+ """Compile frames into a video using H.264 codec."""
+ cmd = [
+ 'ffmpeg', '-y',
+ '-f', 'image2',
+ '-r', str(fps),
+ '-i', f'{frame_dir}/%08d.jpg',
+ '-c:v', 'libx264',
+ '-preset', 'medium',
+ '-crf', '18',
+ '-pix_fmt', 'yuv420p',
+ output_path
+ ]
+ subprocess.run(cmd, check=True)
+ print(f"Successfully compiled video: {output_path}")
+
+def preprocess_videos(video_dir, dataset_name, square_crop=False, fps=24, quality_preset="medium", target_resolution=None):
+ """
+ Preprocess all videos with optional square cropping, customizable FPS, quality, and resolution.
+
+ Args:
+ video_dir (str): Directory containing input videos
+ dataset_name (str): Name of the dataset
+ square_crop (bool): Whether to crop videos to 1:1 aspect ratio (default: True)
+ fps (int): Target frames per second (default: 30)
+ quality_preset (str): Quality preset - "high", "medium", "low", "ultra_low" (default: "medium")
+ target_resolution (int): Target resolution for the shorter side (e.g., 512, 256). None for original
+ """
+ result_dir = ensure_dir(f"../output/{dataset_name}_results")
+
+ # 质量设置
+ quality_settings = {
+ "high": {"qscale": "2", "crf": "18"}, # 高质量
+ "medium": {"qscale": "5", "crf": "23"}, # 中等质量
+ "low": {"qscale": "10", "crf": "28"}, # 低质量
+ "ultra_low": {"qscale": "15", "crf": "35"} # 超低质量
+ }
+
+ current_quality = quality_settings.get(quality_preset, quality_settings["medium"])
+
+ for video_file in os.listdir(video_dir):
+ if not video_file.endswith(".mp4"):
+ continue
+
+ video_name = os.path.splitext(video_file)[0]
+ video_full_path = os.path.join(video_dir, video_file)
+ folder_path = f"{result_dir}/{video_name}"
+
+ frame_path = f"{folder_path}/crop_frame"
+ output_video_path = f"{folder_path}/crop_original_video.mp4"
+
+ # Skip if already processed
+ if os.path.exists(frame_path) and os.listdir(frame_path) and os.path.exists(output_video_path):
+ crop_status = "cropped" if square_crop else "original"
+ print(f"{crop_status.capitalize()} frames and video already exist for {video_name}. Skipping preprocessing.")
+ continue
+
+ try:
+ # Create output directory
+ os.makedirs(frame_path, exist_ok=True)
+
+ # Get video dimensions
+ width, height = get_video_dimensions(video_full_path)
+
+ # 构建视频滤镜
+ filters = []
+
+ if square_crop:
+ # Calculate crop dimensions
+ if width < height:
+ crop_size = width
+ x_offset = 0
+ y_offset = (height - width) // 2
+ else:
+ crop_size = height
+ x_offset = (width - height) // 2
+ y_offset = 0
+ filters.append(f'crop={crop_size}:{crop_size}:{x_offset}:{y_offset}')
+
+ # 添加分辨率缩放
+ if target_resolution:
+ if square_crop:
+ # 方形裁剪后直接缩放到目标分辨率
+ filters.append(f'scale={target_resolution}:{target_resolution}')
+ else:
+ # 保持宽高比缩放
+ filters.append(f'scale=-2:{target_resolution}:force_original_aspect_ratio=decrease')
+
+ # 添加帧率
+ filters.append(f'fps={fps}/1')
+
+ # 组合所有滤镜
+ filter_complex = ','.join(filters)
+
+ # 提取帧的命令
+ cmd = [
+ 'ffmpeg', '-i', video_full_path,
+ '-vf', filter_complex,
+ '-f', 'image2',
+ '-qscale', current_quality["qscale"], # 使用可调节的质量
+ f'{frame_path}/%08d.jpg'
+ ]
+
+ resolution_info = f" (Resolution: {target_resolution})" if target_resolution else ""
+ crop_info = "with square cropping" if square_crop else "without cropping"
+ print(f"Processing {video_file} {crop_info} (FPS: {fps}, Quality: {quality_preset}{resolution_info})")
+
+ subprocess.run(cmd, check=True)
+ print(f"Successfully extracted frames for {video_file}")
+
+ # Compile frames back into a video with optimized settings
+ compile_frames_to_video_optimized(frame_path, output_video_path, fps, quality_preset)
+
+ except Exception as e:
+ print(f"Error preprocessing {video_file}: {str(e)}")
+ continue
+
+def compile_frames_to_video_optimized(frame_dir, output_path, fps=30, quality_preset="medium"):
+ """Compile frames into a video with optimized quality settings."""
+
+ # 质量设置 - CRF值（越高质量越低，文件越小）
+ quality_crf = {
+ "high": "18",
+ "medium": "23",
+ "low": "28",
+ "ultra_low": "35"
+ }
+
+ crf_value = quality_crf.get(quality_preset, "23")
+
+ cmd = [
+ 'ffmpeg', '-y',
+ '-f', 'image2',
+ '-r', str(fps),
+ '-i', f'{frame_dir}/%08d.jpg',
+ '-c:v', 'libx264',
+ '-preset', 'medium', # 可以改为 'fast' 加速编码
+ '-crf', crf_value,
+ '-pix_fmt', 'yuv420p',
+ output_path
+ ]
+ subprocess.run(cmd, check=True)
+ print(f"Successfully compiled optimized video: {output_path} (Quality: {quality_preset})")
+
+# 使用示例：
+
+# 1. 保持原分辨率，降低质量
+# preprocess_videos(video_dir, dataset_name, square_crop=True, fps=30, quality_preset="low")
+
+# 2. 降低分辨率到512x512（方形裁剪）
+# preprocess_videos(video_dir, dataset_name, square_crop=True, fps=30, quality_preset="medium", target_resolution=512)
+
+# 3. 极度压缩：低分辨率 + 超低质量
+# preprocess_videos(video_dir, dataset_name, square_crop=True, fps=30, quality_preset="ultra_low", target_resolution=256)
+
+# 4. 不裁剪，但缩放到较小尺寸
+# preprocess_videos(video_dir, dataset_name, square_crop=False, fps=30, quality_preset="low", target_resolution=480)
+
+def process_npz_files(input_folder_path, output_folder_path):
+ """
+ Process all NPZ files in the specified folder and generate the required output format.
+
+ Args:
+ input_folder_path (str): Path to the folder containing NPZ files
+ output_folder_path (str): Path where output files will be saved
+ """
+ # Get all NPZ files in the folder
+ npz_files = sorted([f for f in os.listdir(input_folder_path) if f.endswith('.npz')])
+ total_frames = len(npz_files)
+
+ output = []
+
+ for idx, npz_file in enumerate(npz_files):
+ file_path = os.path.join(input_folder_path, npz_file)
+ data = np.load(file_path, allow_pickle=True)
+
+ # Process bodies data
+ bodies = data['bodies']
+ body_scores = data['body_scores'][0]
+
+ # Process hands data
+ hands = data['hands']
+ hands_scores = data['hands_scores']
+
+ # Process faces data
+ faces = data['faces'][0]
+ faces_scores = data['faces_scores'][0]
+
+ # Convert coordinates to strings with space separation
+ frame_data = []
+
+ # Add body coordinates and scores
+ for i in range(bodies.shape[0]):
+ frame_data.extend([f"{bodies[i][0]:.8f}", f"{bodies[i][1]:.8f}"])
+ for score in body_scores:
+ frame_data.append(f"{score:.8f}")
+
+ # Add hand coordinates and scores
+ for hand in hands:
+ for point in hand:
+ frame_data.extend([f"{point[0]:.8f}", f"{point[1]:.8f}"])
+ for hand_score in hands_scores:
+ frame_data.extend([f"{score:.8f}" for score in hand_score])
+
+ # Add face coordinates and scores
+ for point in faces:
+ frame_data.extend([f"{point[0]:.8f}", f"{point[1]:.8f}"])
+ for score in faces_scores:
+ frame_data.append(f"{score:.8f}")
+
+ # Add frame count
+ frame_count = idx / (total_frames - 1) if total_frames > 1 else 0
+ frame_data.append(f"{frame_count:.8f}")
+
+ # 验证这一帧的数据点数是否为385
+ if len(frame_data) != 385:
+ print(f"Warning: Frame {idx} in {input_folder_path} has {len(frame_data)} values instead of 385")
+ continue # 跳过这一帧
+
+ # Join all data with spaces
+ output.append(" ".join(frame_data))
+
+ return " ".join(output) + "\n"
\ No newline at end of file
diff --git a/tools-new-2025/main/utils/util.py b/tools-new-2025/main/utils/util.py
new file mode 100644
index 0000000000000000000000000000000000000000..3298b2164e54dae7bc9170e04ac381bdad9d755d
--- /dev/null
+++ b/tools-new-2025/main/utils/util.py
@@ -0,0 +1,128 @@
+import importlib
+import os
+import os.path as osp
+import shutil
+import sys
+from pathlib import Path
+
+import av
+import numpy as np
+import torch
+import torchvision
+from einops import rearrange
+from PIL import Image
+
+
+def seed_everything(seed):
+ import random
+
+ import numpy as np
+
+ torch.manual_seed(seed)
+ torch.cuda.manual_seed_all(seed)
+ np.random.seed(seed % (2**32))
+ random.seed(seed)
+
+
+def import_filename(filename):
+ spec = importlib.util.spec_from_file_location("mymodule", filename)
+ module = importlib.util.module_from_spec(spec)
+ sys.modules[spec.name] = module
+ spec.loader.exec_module(module)
+ return module
+
+
+def delete_additional_ckpt(base_path, num_keep):
+ dirs = []
+ for d in os.listdir(base_path):
+ if d.startswith("checkpoint-"):
+ dirs.append(d)
+ num_tot = len(dirs)
+ if num_tot <