app.py

import streamlit as st
from PIL import Image, ImageDraw
from io import BytesIO
import torch
import torchvision.transforms as transforms
from torchvision.models.detection import fasterrcnn_resnet50_fpn

def draw_bounding_boxes(image, boxes, labels):
    draw = ImageDraw.Draw(image)

    for box, label in zip(boxes, labels):
        draw.rectangle([box[0], box[1], box[2], box[3]], outline="red", width=3)
        draw.text((box[0], box[1]), str(label), fill="red")

    return image

def run_inference(image):
    try:
        # Transform the image
        transform = transforms.Compose([transforms.ToTensor()])
        input_tensor = transform(image).unsqueeze(0)

        # Load a pre-trained Faster R-CNN model
        model = fasterrcnn_resnet50_fpn(pretrained=True)
        model.eval()

        # Perform inference
        with torch.no_grad():
            predictions = model(input_tensor)

        # Extract bounding boxes, labels, and scores
        boxes = predictions[0]['boxes'].cpu().numpy()
        labels = predictions[0]['labels'].cpu().numpy()
        scores = predictions[0]['scores'].cpu().numpy()

        # Apply confidence threshold
        threshold = 0.25
        selected_indices = scores > threshold
        boxes = boxes[selected_indices]
        labels = labels[selected_indices]

        # Draw bounding boxes on the image
        annotated_image = draw_bounding_boxes(image.copy(), boxes.astype(int), labels.astype(int))

        # Save the result to a BytesIO object
        result_bytesio = BytesIO()
        annotated_image.save(result_bytesio, format='JPEG')
        result_bytes = result_bytesio.getvalue()

        # Return the bounding boxes, labels, and result image bytes
        return boxes.tolist(), labels.tolist(), result_bytes
    except Exception as e:
        st.error(f"Error processing the image: {e}")
        return [], [], None

def main():
    st.title("Faster R-CNN Object Detection with Streamlit")

    uploaded_file = st.file_uploader("Choose an image...", type="jpg")

    if uploaded_file is not None:
        # Read the uploaded image
        image = Image.open(uploaded_file)

        # Run inference and get bounding boxes, labels, and result image bytes
        bounding_boxes, labels, result_image_bytes = run_inference(image)

        # Display bounding boxes coordinates
        st.text(f"Bounding Boxes: {bounding_boxes}")

        # Display detected labels
        if labels:
            st.text(f"Detected Labels: {labels}")

        # Display the result image if available
        if result_image_bytes is not None:
            st.image(result_image_bytes, caption="Object Detection Result", use_column_width=True, format="JPEG")

if __name__ == "__main__":
    main()