Outline Introduction Manual Segmentation Thresholding EdgeBased Segmentation RegionBased

Outline � Introduction � Manual Segmentation � Thresholding � Edge-Based Segmentation � Region-Based Segmentation � Deformable Model

Introduction Preprocess Image acquisition, restoration, and enhancement Intermediate process Image segmentation and feature extraction High level process Image interpretation and recognition

Manual Segmentation � The simplest approach � Manual tracing or drawing onto the images

Thresholding � Fast and straightforward T = 102 Before thresholding �������������� After thresholding

Edge-Based Segmentation � Use edge detection algorithms based on Gradient or Laplacian masks � Sobel filter, Robert filter, Prewitt filter etc. Laplacian Mask

Applying the Sobel filter to the CT image

Region-Based Segmentation � Region Growing from Seed Points � Watershed

CT Image after automatic region growing segmentation

Watershed

Example

Deformable Model � Active Contours (Snake)