210 likes | 388 Views
Ch 10 Image Segmentation. Ideally, partition an image into regions corresponding to real world objects. Goals of segmentation. Segments formed by K-means. Segmentation attempted via contour/boundary detection. Clustering versus region-growing. Clustering versus region-growing.
E N D
Ch 10 Image Segmentation Ideally, partition an image into regions corresponding to real world objects. CSE 803 Fall 2008 Stockman
Goals of segmentation CSE 803 Fall 2008 Stockman
Segments formed by K-means CSE 803 Fall 2008 Stockman
Segmentation attempted via contour/boundary detection CSE 803 Fall 2008 Stockman
Clustering versus region-growing CSE 803 Fall 2008 Stockman
Clustering versus region-growing CSE 803 Fall 2008 Stockman
K-means clustering as before: vectors can contain color+texture CSE 803 Fall 2008 Stockman
K-means CSE 803 Fall 2008 Stockman
Histograms can show modes CSE 803 Fall 2008 Stockman
Otsu’s method assumes K=2. It searches for the threshold t that optimizes the intra class variance. CSE 803 Fall 2008 Stockman
Ohlander bifurcated the histogram recursively CSE 803 Fall 2008 Stockman
Recursive histogram-controlled segmentation CSE 803 Fall 2008 Stockman
URL’s of other work • Segmentation lecture http://robots.stanford.edu/cs223b/index.html • Tutorial on graph cut method and assessment of segmentation http://www.cis.upenn.edu/~jshi/GraphTutorial/ CSE 803 Fall 2008 Stockman
Segmentation via region-growing (aggregation) Pixels, or patches, at the lowest level are combined when similar in a hierarchical fashion CSE 803 Fall 2008 Stockman
Decision: combine neighbors? Neighboring pixel or region CSE 803 Fall 2008 Stockman
Aggregation decision CSE 803 Fall 2008 Stockman
Representation of regions CSE 803 Fall 2008 Stockman
Chain codes for boundaries CSE 803 Fall 2008 Stockman
Quad trees divide into quadrants M=mixed; E=empty; F=full CSE 803 Fall 2008 Stockman
Can segment 3D images also • Oct trees subdivide into 8 octants • Same coding: M, E, F used • Software available for doing 3D image processing and differential equations using octree representation. • Can achieve large compression factor. CSE 803 Fall 2008 Stockman
More URLs of other work • Mean shift description http://cmp.felk.cvut.cz/cmp/courses/ZS1/slidy/meanShiftSeg.pdf CSE 803 Fall 2008 Stockman