Add files using upload-large-folder tool

README.md

@@ -2,61 +2,96 @@
 pretty_name: SegSTRONG-C
 license: cc-by-nc-sa-4.0
 viewer: false
+task_categories:
+- image-segmentation
+size_categories:
+- 10K<n<100K
 tags:
 - surgery
 - medical-imaging
-- segmentation
+- surgical-tool-segmentation
 - robustness
+- corruption
 - stereo
 - endovis
-task_categories:
-- image-segmentation
+- miccai
 ---
 
 # SegSTRONG-C
 
-**Segmenting Surgical Tools Robustly On Non-adversarial Generated Corruptions**, the EndoVis'24 challenge (MICCAI 2024), organized by the ARCADE Lab at Johns Hopkins University.
+**Segmenting Surgical Tools Robustly On Non-adversarial Generated Corruptions**, the EndoVis'24 challenge at MICCAI 2024, organized by the ARCADE Lab at Johns Hopkins University.
 
-- Website: https://segstrongc.cs.jhu.edu/
-- Paper: https://arxiv.org/abs/2407.11906
-- Code and baselines: https://github.com/hding2455/CaRTS
+[Challenge website](https://segstrongc.cs.jhu.edu/) · [Paper (arXiv:2407.11906)](https://arxiv.org/abs/2407.11906) · [Code and baselines (CaRTS)](https://github.com/hding2455/CaRTS)
 
-The task is binary segmentation of robot-assisted surgical tools against the tissue background, in stereo endoscopic images that can be treated as video sequences. The challenge asks how well models trained on clean data hold up under realistic corruptions.
+## Overview
 
-The data is paired by robotic replay: the same tool motion is replayed under different conditions (clean, smoke, bleeding, low brightness, alternative background). Every condition of a subsequence therefore shares one ground-truth mask per frame.
+SegSTRONG-C asks how robust surgical tool segmentation is to realistic, non-adversarial corruptions. The task is binary segmentation of robot-assisted surgical tools against the tissue background, in stereo endoscopic images that can also be treated as video.
+
+The data is paired by robotic replay: the same tool motion is recorded clean and again under smoke, bleeding, low brightness and an alternative background. Every condition of a subsequence therefore shows the same instant at each frame index and shares one ground-truth mask. Models train on clean data only and are evaluated on the corrupted versions.
 
 ## Contents
 
-| Split | Sequences / subsequences | Frames per subsequence | Conditions | Images per condition |
-|---|---|---|---|---|
-| train | 3/0, 3/2, 4/0, 4/1, 4/2, 5/0, 5/2, 7/0, 7/1, 8/1, 8/2 (11) | 300 | regular | 3,300 left + 3,300 right |
-| test | 9/0, 9/1, 9/2 (3) | 300 | regular, smoke, blood, low_brightness, bg_change | 900 left + 900 right each |
-| val | *being added; the Google Drive copy was rate-limited at upload time* | | | |
+| Split | Sequence / subsequences | Conditions | Stereo frames | Images (incl. masks) | Size |
+|---|---|---|---|---|---|
+| train | 3/0, 3/2, 4/0, 4/1, 4/2, 5/0, 5/2, 7/0, 7/1, 8/1, 8/2 | regular | 3,300 | 13,200 | 22.5 GB |
+| val | 1/0, 1/1, 1/2 | regular, smoke, blood, low_brightness, bg_change | 900 per condition | 10,800 | 30.1 GB |
+| test | 9/0, 9/1, 9/2 | regular, smoke, blood, low_brightness, bg_change | 900 per condition | 10,800 | 29.9 GB |
+
+Every subsequence has 300 frames, each with a left and a right view.
+
+The challenge used part of the val and test folders:
+- **Validation:** `regular` and `bg_change`. That's 3,600 frames: 3 × 300 × 2 views × 2 conditions.
+- **Ranked test:** `smoke`, `blood` and `low_brightness`. That's 5,400 frames: 3 × 300 × 2 views × 3 conditions.
 
-On the challenge website, training is 6,600 frames (11 subsequences × 300 × 2 views), and testing is 5,400 frames (3 × 300 × 2 views × the three ranked corruptions: smoke, bleeding, low brightness). The test folders also hold the clean `regular` frames and `bg_change`.
+The release includes every condition for both splits.
 
 ## Layout
 
 ```
 SegSTRONGC_train/<seq>/<subseq>/regular/{left,right}/<frame>.png
 SegSTRONGC_train/<seq>/<subseq>/ground_truth/{left,right}/<frame>.png
+SegSTRONGC_val/val/<seq>/<subseq>/{regular,smoke,blood,low_brightness,bg_change}/{left,right}/<frame>.png
+SegSTRONGC_val/val/<seq>/<subseq>/ground_truth/{left,right}/<frame>.png
 SegSTRONGC_test/test/<seq>/<subseq>/{regular,smoke,blood,low_brightness,bg_change}/{left,right}/<frame>.png
 SegSTRONGC_test/test/<seq>/<subseq>/ground_truth/{left,right}/<frame>.png
 ```
 
-- Images: RGB PNG, 1920×1080, named `0.png` … `299.png`.
-- Masks (`ground_truth`): 8-bit grayscale PNG, 1920×1080, **0 = background, 255 = tool**. The website describes the labels as 1 = tool and 0 = background, so divide by 255.
-- The same `<frame>.png` in every condition folder of a subsequence shows the same instant and uses the same mask.
+- **Images:** RGB PNG, 1920×1080, named `0.png` … `299.png`.
+- **Masks** (`ground_truth`): 8-bit grayscale PNG, 1920×1080, **0 = background and 255 = tool**, so divide by 255 to get labels. Every mask contains some tool; tools average 13–20% of pixels, depending on the split.
+- **Pairing:** `<frame>.png` in any condition folder of a subsequence pairs with `ground_truth/<side>/<frame>.png` of the same subsequence.
 
-This is the original release unchanged, unzipped from `SegSTRONGC_train.zip` and `SegSTRONGC_test.zip`.
+The files are the original release, unzipped from `SegSTRONGC_{train,val,test}.zip` with the folder structure unchanged.
+
+## Quick start
+
+```python
+from huggingface_hub import snapshot_download
+import numpy as np
+from PIL import Image
+
+# Download one split only (the full dataset is about 82 GB).
+root = snapshot_download("hding2455/SegSTRONGC", repo_type="dataset",
+                         allow_patterns=["SegSTRONGC_train/*"])
+
+frame = f"{root}/SegSTRONGC_train/3/0"
+image = np.array(Image.open(f"{frame}/regular/left/0.png"))           # (1080, 1920, 3) uint8
+mask = np.array(Image.open(f"{frame}/ground_truth/left/0.png")) > 0   # (1080, 1920) bool, True = tool
+```
+
+For corrupted evaluation data, use for example `allow_patterns=["SegSTRONGC_test/test/9/*/smoke/*", "SegSTRONGC_test/test/9/*/ground_truth/*"]`.
 
 ## Evaluation
 
-The challenge ranks entries by the Dice similarity coefficient (DSC) and normalized surface distance (NSD) at several tolerances, aggregated over the three corruption domains.
+The challenge ranked entries by the Dice similarity coefficient (DSC) and normalized surface distance (NSD) at several tolerances, aggregated over the three corruption domains: smoke, bleeding and low brightness. Baseline models and evaluation code are in [CaRTS](https://github.com/hding2455/CaRTS).
 
 ## License
 
-The SegSTRONG-C data is released under the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.0)](https://creativecommons.org/licenses/by-nc-sa/4.0/): use it with attribution (cite the paper below), for non-commercial purposes only, and share derivatives under the same license. The CaRTS code is licensed separately (2-clause BSD for academic use).
+The SegSTRONG-C data is released under the [Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International license (CC BY-NC-SA 4.0)](https://creativecommons.org/licenses/by-nc-sa/4.0/):
+- use it with attribution (cite the paper below);
+- for non-commercial purposes only;
+- share derivatives under the same license.
+
+The CaRTS code is licensed separately.
 
 ## Citation