Image Segmentation

This project represents an image as a fully connected, non-directional graph and partitions it into segments based on common characteristics like color or intensity.

Implemented techniques:

• Spectral Clustering: Partitions into k clusters using graph Laplacian eigenvalues (Algorithm)
• Normalized Cuts: Segments using the k smallest eigenvalues for graph partitioning (Algorithm)
• Recursive Normalized Cuts: Automatically determines the number of clusters using the Ncut(A, B) metric (Algorithm)