Module 4: Feature Extraction and Matching

1. Learning Objectives

• Understand how to detect and describe local image features
• Compare feature detectors like Harris, FAST, SIFT, SURF, and ORB
• Apply feature matching techniques using OpenCV
• Use feature matches for image alignment and recognition

2. Key Concepts

• Keypoints: Interest points in the image that are repeatable and distinctive (e.g., corners, blobs)
• Descriptors: Vectors that describe the region around a keypoint
• Feature Matching: Comparing descriptors across images to find correspondences

3. Keypoint Detectors

• Harris Corner Detector: Detects corners using gradient-based scoring
• FAST (Features from Accelerated Segment Test): Efficient for real-time keypoint detection

4. Descriptors

• SIFT: Scale-Invariant Feature Transform (good for scale/rotation changes)
• SURF: Speeded-Up Robust Features (patented)
• ORB: Oriented FAST and Rotated BRIEF (efficient and open-source)

5. Feature Matching Techniques

• Brute Force Matcher: Compares descriptors one-by-one
• FLANN (Fast Library for Approximate Nearest Neighbors): Faster for large datasets

6. Hands-On Lab

Use OpenCV to detect keypoints and match features between two uploaded images.

import cv2
import numpy as np
from google.colab.patches import cv2_imshow
from google.colab import files

# Upload two images
uploaded = files.upload()
img1 = cv2.imread(list(uploaded.keys())[0], 0)  # Grayscale
uploaded = files.upload()
img2 = cv2.imread(list(uploaded.keys())[0], 0)

# Detect ORB keypoints and descriptors
orb = cv2.ORB_create()
kp1, des1 = orb.detectAndCompute(img1, None)
kp2, des2 = orb.detectAndCompute(img2, None)

# Match descriptors using BFMatcher
bf = cv2.BFMatcher(cv2.NORM_HAMMING, crossCheck=True)
matches = bf.match(des1, des2)
matches = sorted(matches, key=lambda x: x.distance)

# Draw top 20 matches
matched_img = cv2.drawMatches(img1, kp1, img2, kp2, matches[:20], None, flags=2)
cv2_imshow(matched_img)

7. Real-World Applications

• Panorama stitching
• Object recognition and tracking
• Augmented reality anchor detection

8. Assignment