Template Matching is a core computer vision technique for finding objects of interest in an image. 

This video will discuss three main commonly used template-matching algorithms and their pros and cons. 
- Zero-mean filtering
- Sum of Squared Distances (SSD)
- Normalized Cross-Correlation (NCC)

Video made with Manim: https://www.manim.community/

Jia-Bin Huang

An accessible walkthrough of template matching techniques in computer vision. Covers cross-correlation with zero-mean filtering, sum of squared differences (SSD), and normalized cross-correlation (NCC). Explains why naive cross-correlation fails by highlighting bright regions, how subtracting mean intensity fixes it, and how NCC handles varying lighting conditions. Compares the three methods on speed vs. accuracy vs. robustness trade-offs.

How Machines Find Patterns [Template Matching]