app.py

import numpy as np
import cv2 as cv
import gradio as gr

def match_features(target_img, comp_img1, comp_img2, comp_img3, comp_img4, comp_img5, comp_img6, comp_img7, comp_img8):
    # Initialize list to store results
    result_images = []
    match_counts = []
    
    # List of comparison images
    comparison_imgs = [comp_img1, comp_img2, comp_img3, comp_img4, comp_img5, comp_img6, comp_img7, comp_img8]
    
    # Convert target image to grayscale OpenCV format
    target_cv = np.array(target_img.convert("L"))
    
    # Extract SIFT features from target image
    sift = cv.SIFT_create()
    kp_target, des_target = sift.detectAndCompute(target_cv, None)
    
    # Initialize FLANN matcher
    FLANN_INDEX_KDTREE = 1
    index_params = dict(algorithm=FLANN_INDEX_KDTREE, trees=5)
    search_params = dict(checks=50)
    flann = cv.FlannBasedMatcher(index_params, search_params)
    
    # Process each comparison image
    for img in comparison_imgs:
        # Create a default blank image with "No image" text
        blank_img = np.zeros((400, 800, 3), dtype=np.uint8)
        cv.putText(blank_img, "No image provided", (250, 200), cv.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 2)
        
        if img is None:
            result_images.append(blank_img)
            match_counts.append(0)
            continue
            
        # Convert to grayscale OpenCV format
        img_cv = np.array(img.convert("L"))
        
        # Extract SIFT features
        kp_img, des_img = sift.detectAndCompute(img_cv, None)
        
        # Skip if no features detected
        if des_img is None or des_target is None:
            cv.putText(blank_img, "No features detected", (250, 200), cv.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 2)
            result_images.append(blank_img)
            match_counts.append(0)
            continue
        
        try:
            # Match features with FLANN
            matches = flann.knnMatch(des_target, des_img, k=2)
            
            # Apply ratio test
            matchesMask = [[0, 0] for i in range(len(matches))]
            good_matches = []
            for i, pair in enumerate(matches):
                if len(pair) < 2:
                    continue
                m, n = pair
                if m.distance < 0.7 * n.distance:
                    matchesMask[i] = [1, 0]
                    good_matches.append(m)
            
            # Count good matches
            match_count = len(good_matches)
            match_counts.append(match_count)
            
            # Draw matches
            draw_params = dict(
                matchColor=(0, 255, 0),
                singlePointColor=(255, 0, 0),
                matchesMask=matchesMask,
                flags=cv.DrawMatchesFlags_DEFAULT
            )
            
            result_img = cv.drawMatchesKnn(target_cv, kp_target, img_cv, kp_img, matches, None, **draw_params)
            
            # Convert to RGB for display
            result_img = cv.cvtColor(result_img, cv.COLOR_BGR2RGB)
            
            # Add match count text
            font = cv.FONT_HERSHEY_SIMPLEX
            font_scale = 1
            font_color = (255, 255, 255)
            thickness = 2
            
            h, w = result_img.shape[:2]
            text = f"Matches: {match_count}"
            
            (text_width, text_height), _ = cv.getTextSize(text, font, font_scale, thickness)
            
            x = (w - text_width) // 2
            y = h - 20
            
            cv.rectangle(result_img, (x-5, y-text_height-5), (x+text_width+5, y+5), (0,0,0), -1)
            cv.putText(result_img, text, (x, y), font, font_scale, font_color, thickness)
            
            result_images.append(result_img)
        except Exception as e:
            # Handle any errors
            error_img = np.zeros((400, 800, 3), dtype=np.uint8)
            cv.putText(error_img, f"Error: {str(e)}", (50, 200), cv.FONT_HERSHEY_SIMPLEX, 0.7, (255, 255, 255), 1)
            result_images.append(error_img)
            match_counts.append(0)
    
    # Ensure we have 8 results
    while len(result_images) < 8:
        blank_img = np.zeros((400, 800, 3), dtype=np.uint8)
        cv.putText(blank_img, "No image provided", (250, 200), cv.FONT_HERSHEY_SIMPLEX, 1, (255, 255, 255), 2)
        result_images.append(blank_img)
        match_counts.append(0)
    
    # Return all match results and counts
    return result_images + [match_counts]

# Create Gradio interface with 1 target and up to 8 comparison images
with gr.Blocks(title="Image Feature Matching Comparison") as iface:
    gr.Markdown("# Image Feature Matching with SIFT+FLANN")
    gr.Markdown("""
    Upload a target image and up to 8 comparison images to find feature matches using 
    SIFT (Scale-Invariant Feature Transform) with FLANN (Fast Library for Approximate Nearest Neighbors).
    The number of matches will be displayed for each comparison.
    """)
    
    with gr.Row():
        target_input = gr.Image(type="pil", label="Target Image")
    
    with gr.Row():
        comp_img1 = gr.Image(type="pil", label="Comparison Image 1")
        comp_img2 = gr.Image(type="pil", label="Comparison Image 2")
    
    with gr.Row():
        comp_img3 = gr.Image(type="pil", label="Comparison Image 3")
        comp_img4 = gr.Image(type="pil", label="Comparison Image 4")
    
    with gr.Row():
        comp_img5 = gr.Image(type="pil", label="Comparison Image 5")
        comp_img6 = gr.Image(type="pil", label="Comparison Image 6")
    
    with gr.Row():
        comp_img7 = gr.Image(type="pil", label="Comparison Image 7")
        comp_img8 = gr.Image(type="pil", label="Comparison Image 8")
    
    compare_btn = gr.Button("Compare Images")
    
    with gr.Row():
        result1 = gr.Image(label="Result 1")
        result2 = gr.Image(label="Result 2")
    
    with gr.Row():
        result3 = gr.Image(label="Result 3")
        result4 = gr.Image(label="Result 4")
    
    with gr.Row():
        result5 = gr.Image(label="Result 5")
        result6 = gr.Image(label="Result 6")
    
    with gr.Row():
        result7 = gr.Image(label="Result 7")
        result8 = gr.Image(label="Result 8")
    
    match_counts_output = gr.JSON(label="Match Counts")
    
    compare_btn.click(
        fn=match_features,
        inputs=[
            target_input,
            comp_img1, comp_img2, comp_img3, comp_img4,
            comp_img5, comp_img6, comp_img7, comp_img8
        ],
        outputs=[
            result1, result2, result3, result4,
            result5, result6, result7, result8,
            match_counts_output
        ]
    )

# Launch the interface
iface.launch()