Image Segmentation
FleXray
ONNX
Safetensors
PyTorch
medical-image-segmentation
x-ray
radiograph
anatomy
Instructions to use VictorButoi/flexray with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- FleXray
How to use VictorButoi/flexray with FleXray:
pip install flexray
from fxr.inference import FleXraySegmenter segmenter = FleXraySegmenter.from_pretrained("VictorButoi/flexray") prediction = segmenter.predict("image.png", threshold=0.5) masks = prediction.masks - Notebooks
- Google Colab
- Kaggle
Point website links to flexray.csail.mit.edu
Browse files
README.md
CHANGED
|
@@ -15,7 +15,7 @@ tags:
|
|
| 15 |
|
| 16 |
# FleXray: `VictorButoi/flexray`
|
| 17 |
|
| 18 |
-
- Website and in-browser demo: [
|
| 19 |
- Code: [github.com/VictorButoi/FleXray](https://github.com/VictorButoi/FleXray)
|
| 20 |
- Data: [`VictorButoi/flexray-data`](https://huggingface.co/datasets/VictorButoi/flexray-data)
|
| 21 |
- Tutorial: [Colab notebook](https://colab.research.google.com/drive/1MNIeN9LN-tY8wAifWSolxmFZi3Q0BlVX)
|
|
@@ -194,7 +194,7 @@ RAM-W600, VinDr-Rib), spanning lungs, ribs, peripheral bones, spine, and
|
|
| 194 |
pelvis. Against generalist baselines (FluoroSAM, TotalSegmentator2D, PAXray)
|
| 195 |
it achieves the highest macro Dice in every comparison (9 of 9). Per-dataset
|
| 196 |
numbers and confidence intervals are in the paper; the benchmark figure is on
|
| 197 |
-
the [project website](https://
|
| 198 |
Evaluation ignores ground-truth labels covering less than 0.1% of the image.
|
| 199 |
|
| 200 |
## Intended use and limitations
|
|
|
|
| 15 |
|
| 16 |
# FleXray: `VictorButoi/flexray`
|
| 17 |
|
| 18 |
+
- Website and in-browser demo: [flexray.csail.mit.edu](https://flexray.csail.mit.edu/)
|
| 19 |
- Code: [github.com/VictorButoi/FleXray](https://github.com/VictorButoi/FleXray)
|
| 20 |
- Data: [`VictorButoi/flexray-data`](https://huggingface.co/datasets/VictorButoi/flexray-data)
|
| 21 |
- Tutorial: [Colab notebook](https://colab.research.google.com/drive/1MNIeN9LN-tY8wAifWSolxmFZi3Q0BlVX)
|
|
|
|
| 194 |
pelvis. Against generalist baselines (FluoroSAM, TotalSegmentator2D, PAXray)
|
| 195 |
it achieves the highest macro Dice in every comparison (9 of 9). Per-dataset
|
| 196 |
numbers and confidence intervals are in the paper; the benchmark figure is on
|
| 197 |
+
the [project website](https://flexray.csail.mit.edu/#results).
|
| 198 |
Evaluation ignores ground-truth labels covering less than 0.1% of the image.
|
| 199 |
|
| 200 |
## Intended use and limitations
|