Corner Detection and HOG Feature Comparison

Objectives:

1. Gaussian Noise and Smoothing
2. Harris Corner Detection
3. Non-maximum Suppression
4. Corner Localization
5. Extract Feature Vectors with HOG
6. Feature Comparison

Interactively controlled parameters:

• Variance for gaussian (scale)
• Smoothing value for gaussian smoothing (filter size)
• Neighborhood size for computing the correlation matrix
• Weight of the trace in the Harris corner detection
• Threshold value for non-maximum Suppression

Program was implemented using Python and OpenCV. Refer the report for further implementation details and instructions to run the code: View Report

Results:

1. Corner Detection with implemented Harris corner detection function and OpenCV:

Implemented Harris Corner Detection	OpenCV

2. Corner Detection with Gaussian Noise (variance = 1) and Gaussian Smoothing (filter size = 15):

Implemented Harris Corner Detection	OpenCV

3. Feature Comparison with Features Extracted:

Image 1	Image 2