Region Segmentation Lecture

Abstract

This lecture overviews Region Segmentation and Texture Description that have many applications in in image analysis, notably in shape description and object recognition It covers the following topics in detail: Image segmentation by thresholding. Region growing techniques. Split techniques and Quadtrees. Split /merge techniques. Relaxation algorithms in region analysis. Neural region segmentation.

Connected component labeling for object counting and object area measurements. Various Texture Description methods, e.g., histogram-based features, histogram of grey-level differences, Run-length techniques, co-occurrence matrices,  Spectral texture characterization.

Street scene segmentation.

Region-Segmentation-v3.4

Understanding Questionnaire

https://docs.google.com/forms/region-segmentation

Programming Exercise

1. Region Segmentation