alex17cmbs/chexnet-pneumonia

🫁 densenet121-chexnet-binary - NIH Chest X-ray Classification

Ce modèle a été entraîné pour la classification multi-label de pathologies thoraciques à partir de radiographies X-ray.

📋 Description

• Architecture: densenet121-chexnet-binary
• Tâche: Classification multi-label (15 classes)
• Dataset: NIH Chest X-ray
• Framework: PyTorch

🏷️ Classes

ID	Pathologie
0	No Finding
1	Atelectasis
2	Cardiomegaly
3	Effusion
4	Infiltration
5	Mass
6	Nodule
7	Pneumonia
8	Pneumothorax
9	Consolidation
10	Edema
11	Emphysema
12	Fibrosis
13	Pleural_Thickening
14	Hernia

📊 Performance

Metric	Value
AUC-ROC	0.6079
F1-Score	0.0609
Precision	0.0328
Recall	0.4288
Accuracy	0.7130
Specificity	0.7193

🚀 Utilisation

import torch
from src.models import create_model

# Charger le modèle
model = create_model('densenet121-chexnet-binary', num_classes=15, pretrained=False)
model.load_state_dict(torch.load('pytorch_model.bin'))
model.eval()

# Prédiction
from torchvision import transforms
from PIL import Image

transform = transforms.Compose([
    transforms.Resize((224, 224)),
    transforms.ToTensor(),
    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
])

image = Image.open('chest_xray.png').convert('RGB')
input_tensor = transform(image).unsqueeze(0)

with torch.no_grad():
    logits = model(input_tensor)
    probs = torch.sigmoid(logits)
    
print(probs)

⚙️ Training Configuration

{
  "num_epochs": 30,
  "learning_rate": 0.001,
  "weight_decay": 1e-05,
  "early_stopping_patience": 30
}

📚 Citation

@inproceedings{Wang_2017,
    title = {ChestX-Ray8: Hospital-Scale Chest X-Ray Database and Benchmarks},
    author = {Wang, Xiaosong and others},
    booktitle = {IEEE CVPR},
    year = {2017}
}

📄 License

MIT License