Traditional Image Processing for Filtering, 3D Reconstruction, and Image Stitching

Objectives

You will implement and compare classical computer vision techniques for:

1. Image filtering and enhancement
2. 3D reconstruction from stereo images
3. Image stitching from multiple views

Part A – Image Filtering

Task 1: Source code

Apply and compare different traditional filters to remove noise and enhance images:

• Mean filter
• Gaussian filter
• Median filter
• Laplacian sharpening

Deliverables:

• Original, noisy, and filtered images (side-by-side)
• Comparison of filters on noise reduction and edge preservation

Task 2: Comparative Analysis

Compare at least two alternative methods or parameter settings. Discuss:

• Strengths and weaknesses
• Quantitative evaluation (e.g., PSNR, number of inliers)
• Qualitative visual inspection

Task 3: Final Report Guidelines

Submit a clearly written Markdown report including:

• Introduction and motivation
• Description of methods with equations/diagrams
• Implementation steps and results
• Comparison and analysis
• Conclusion and possible extensions