Views
No views yet
1├── Data Collection and Preprocessing/
2│ ├── Data_collection_Mimic.ipynb
3│ ├── Data_preprocess.ipynb
4│ ├── Radgraph Based Report Cleaning.ipynb
5│ └── train_data_json_gen.ipynb
6│
7├── Evaluate/
8│ ├── Evaluate.ipynb
9│ └── Results_IU_Xray/ # Contains evaluation results on IU X-ray dataset
10│
11├── llava_phi/
12│ ├── Dual Slava train.ipynb # Training pipeline
13│ └── generation.ipynb # Inference/report generation
14│
15├── requirements.txt
16└── README.md| Dataset | BLEU | ROUGE-L | METEOR | BERT | RadGraph F1 | CheXbert F1 |
|---|---|---|---|---|---|---|
| MIMIC-CXR | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
| IU X-Ray | ✅ | ✅ | ✅ | ✅ | ✅ | ✅ |
/Evaluate/Results_IU_Xray)1# Clone repo
2git clone https://github.com/Clintonkjkj/Dual-View-Slava-CXR.git
3cd Dual-View-Slava-CXR
4
5# Set up virtual environment
6python -m venv venv
7source venv/bin/activate # or venv\Scripts\activate
8
9# Install dependencies
10pip install -r requirements.txtllava_phi/Dual Slava train.ipynb after preparing data using:Data_collection_Mimic.ipynbData_preprocess.ipynbRadgraph Based Report Cleaning.ipynbtrain_data_json_gen.ipynbllava_phi/generation.ipynb with both frontal and lateral views, plus a prompt (e.g., "Generate a radiology report").
1@misc{dualviewslava2025,
2 title={Dual View SLaVA-CXR: Structured Radiology Reporting via Multi-View Chest X-rays},
3 author={Clinton KJ et al.},
4 year={2025},
5 note={Capstone Project}
6}