Spaces:
Sleeping
Sleeping
update readme.md
Browse filesupdated readme.md
README.md
CHANGED
@@ -1,13 +1,63 @@
|
|
1 |
-
|
2 |
-
|
3 |
-
|
4 |
-
|
5 |
-
|
6 |
-
|
7 |
-
|
8 |
-
|
9 |
-
|
10 |
-
|
11 |
-
|
12 |
-
|
13 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
# Medical VLM with SAM-2 and CheXagent
|
2 |
+
|
3 |
+
A comprehensive medical imaging analysis tool that combines:
|
4 |
+
- Qwen-VLM for medical visual question answering
|
5 |
+
- SAM-2 (Segment Anything Model 2) for automatic medical image segmentation
|
6 |
+
- CheXagent for structured chest X-ray report generation
|
7 |
+
|
8 |
+
## Features
|
9 |
+
|
10 |
+
1. **Medical Q&A**: Ask questions about medical images using the Qwen-VLM model
|
11 |
+
2. **Automatic Masking**: Segment medical images automatically using SAM-2
|
12 |
+
3. **Structured Report Generation**: Generate detailed chest X-ray reports using CheXagent
|
13 |
+
4. **Visual Grounding**: Locate specific findings in medical images
|
14 |
+
|
15 |
+
## Setup
|
16 |
+
|
17 |
+
1. Clone the repository:
|
18 |
+
```bash
|
19 |
+
git clone https://github.com/pascal-maker/medicalvlm.git
|
20 |
+
cd medicalvlm
|
21 |
+
```
|
22 |
+
|
23 |
+
2. Create and activate a virtual environment:
|
24 |
+
```bash
|
25 |
+
python -m venv chexagent_env
|
26 |
+
source chexagent_env/bin/activate # On Windows: chexagent_env\Scripts\activate
|
27 |
+
```
|
28 |
+
|
29 |
+
3. Install dependencies:
|
30 |
+
```bash
|
31 |
+
pip install -r requirements.txt
|
32 |
+
```
|
33 |
+
|
34 |
+
4. Download required model checkpoints:
|
35 |
+
- SAM-2 checkpoint: Place in `checkpoints/sam2.1_hiera_large.pt`
|
36 |
+
- Other model weights will be downloaded automatically on first run
|
37 |
+
|
38 |
+
## Usage
|
39 |
+
|
40 |
+
Run the Gradio interface:
|
41 |
+
```bash
|
42 |
+
python app.py
|
43 |
+
```
|
44 |
+
|
45 |
+
The web interface will be available at `http://localhost:7860`
|
46 |
+
|
47 |
+
## Requirements
|
48 |
+
|
49 |
+
- Python 3.8+
|
50 |
+
- PyTorch
|
51 |
+
- CUDA-compatible GPU (recommended)
|
52 |
+
- See `requirements.txt` for full list of dependencies
|
53 |
+
|
54 |
+
## License
|
55 |
+
|
56 |
+
[Your chosen license]
|
57 |
+
|
58 |
+
## Acknowledgments
|
59 |
+
|
60 |
+
- [Qwen-VLM](https://github.com/QwenLM/Qwen-VL)
|
61 |
+
- [SAM-2](https://github.com/facebookresearch/segment-anything)
|
62 |
+
- [CheXagent](https://github.com/stanfordmlgroup/CheXagent)
|
63 |
+
|