Create app.py

app.py

@@ -0,0 +1,154 @@
+import gradio as gr
+import torch
+import torch.nn as nn
+import torchvision.transforms as transforms
+from efficientnet_pytorch import EfficientNet
+from PIL import Image
+import numpy as np
+import os
+from fpdf import FPDF
+from datetime import datetime
+
+# Define HardSwish activation
+class HardSwish(nn.Module):
+    def __init__(self):
+        super(HardSwish, self).__init__()
+
+    def forward(self, x):
+        return x * (torch.clamp(x + 3, 0, 6) / 6)
+
+# Define model class
+class CustomEfficientNet(nn.Module):
+    def __init__(self, num_classes):
+        super(CustomEfficientNet, self).__init__()
+        self.model = EfficientNet.from_name('efficientnet-b3')
+        num_ftrs = self.model._fc.in_features
+        self.model._fc = nn.Sequential(
+            nn.Linear(num_ftrs, 512),
+            HardSwish(),
+            nn.Dropout(p=0.4),
+            nn.Linear(512, num_classes)
+        )
+
+    def forward(self, x):
+        return self.model(x)
+
+# Class names
+class_names = ['No Finding', 'Enlarged Cardiomediastinum', 'Cardiomegaly', 'Lung Opacity',
+               'Lung Lesion', 'Edema', 'Consolidation', 'Pneumonia', 'Atelectasis', 'Pneumothorax',
+               'Pleural Effusion', 'Pleural Other', 'Fracture', 'Support Devices']
+
+# Device config
+device = torch.device('cuda' if torch.cuda.is_available() else 'cpu')
+
+# Load the model
+model = CustomEfficientNet(num_classes=14)
+checkpoint = torch.load('Final_global_model.pth.tar', map_location=device)
+if 'state_dict' in checkpoint:
+    model.load_state_dict(checkpoint['state_dict'])
+else:
+    model.load_state_dict(checkpoint)
+model = model.to(device)
+model.eval()
+
+# Transform
+transform = transforms.Compose([
+    transforms.Resize((300, 300)),
+    transforms.ToTensor(),
+    transforms.Normalize([0.485, 0.456, 0.406], [0.229, 0.224, 0.225])
+])
+
+# Prediction and PDF generation
+def predict_and_generate_pdf(image, patient_name, xray_date):
+    if image is None:
+        raise ValueError("❌ Error: No image uploaded.")
+
+    if not isinstance(image, Image.Image):
+        image = Image.fromarray(image)
+
+    if image.mode != 'RGB':
+        image = image.convert('RGB')
+
+    # Preprocess
+    img = transform(image).unsqueeze(0).to(device)
+
+    # Predict
+    with torch.no_grad():
+        outputs = model(img)
+        probs = torch.sigmoid(outputs).cpu().numpy()[0]
+
+    # Process results
+    results = {class_names[i]: float(probs[i]) for i in range(len(class_names))}
+    sorted_results = dict(sorted(results.items(), key=lambda item: item[1], reverse=True))
+    top5 = list(sorted_results.items())[:5]
+
+    # Generate Comment
+    top_label, top_prob = top5[0]
+    if top_label == "No Finding" and top_prob > 0.5:
+        comment = "No major abnormal findings detected."
+    elif top_prob > 0.5:
+        comment = f"High likelihood of {top_label}."
+    else:
+        comment = f"Possibility of {top_label}, but with low confidence."
+
+    # Save thumbnail
+    image_path = "xray_temp.jpg"
+    image.thumbnail((200, 200))
+    image.save(image_path)
+
+    # Create PDF
+    pdf = FPDF()
+    pdf.add_page()
+    pdf.set_font("Arial", 'B', 16)
+    pdf.cell(0, 10, "Chest X-ray Report", ln=True, align='C')
+    pdf.ln(10)
+
+    pdf.set_font("Arial", '', 12)
+    pdf.cell(0, 10, f"Patient Name: {patient_name}", ln=True)
+    pdf.cell(0, 10, f"X-ray Date: {xray_date}", ln=True)
+    pdf.cell(0, 10, f"Report Generated: {datetime.now().strftime('%Y-%m-%d %H:%M:%S')}", ln=True)
+    pdf.ln(10)
+
+    pdf.image(image_path, w=80)
+    pdf.ln(10)
+
+    pdf.set_font("Arial", 'B', 14)
+    pdf.cell(0, 10, "Top 5 Predictions:", ln=True)
+    pdf.set_font("Arial", '', 12)
+    for disease, prob in top5:
+        pdf.cell(0, 10, f"{disease}: {prob*100:.2f}%", ln=True)
+
+    pdf.ln(10)
+    pdf.set_font("Arial", 'B', 14)
+    pdf.cell(0, 10, "Comment:", ln=True)
+    pdf.set_font("Arial", '', 12)
+    pdf.multi_cell(0, 10, comment)
+
+    # Save PDF
+    pdf_output_path = "report.pdf"
+    pdf.output(pdf_output_path)
+
+    return pdf_output_path
+
+# Gradio Interface
+with gr.Blocks(theme="default") as demo:
+    gr.Markdown("# 🩺 Chest X-ray Disease Classification")
+    gr.Markdown("Upload a chest X-ray, enter patient's name and date, and download a PDF report.")
+
+    with gr.Row():
+        with gr.Column():
+            image_input = gr.Image(type="pil", label="Upload Chest X-ray")
+            name_input = gr.Textbox(label="Patient Name")
+            date_input = gr.Textbox(label="X-ray Date (YYYY-MM-DD)", placeholder="2025-04-27")
+            submit_btn = gr.Button("Analyze & Generate Report")
+        with gr.Column():
+            file_output = gr.File(label="Download Report (PDF)")
+
+    submit_btn.click(
+        fn=predict_and_generate_pdf,
+        inputs=[image_input, name_input, date_input],
+        outputs=file_output
+    )
+
+if __name__ == "__main__":
+    demo.launch()