Region Detection Defining regions of an image
Introduction All pixels belong to a region Object Part of object Background Find region Constituent pixels Boundary
Region Detection A set of pixels P An homogeneity predicate H(P) Partition P into regions {R}, such that
Point based methods – thresholding If Regions are different brightness or colour Then Can be differentiated using this
Global thresholds Compute threshold from whole image Incorrect in some regions
Local thresholds Divide image into regions Compute threshold per region Merge thresholds across region boundaries
Region Growing All pixels belong to a region Select a pixel Grow the surrounding region
Slow Algorithm If a pixel is Not assigned to a region Adjacent to region Has colour properties not different to region’s Then Add to region Update region properties
Split and Merge Initialise image as a region While region is not homogeneous Split into quadrants and examine homogeneity
Recursive Splitting Split(P) { If (!H(P)) { P subregions 1 … 4; Split (subregion 1); Split (subregion 2); Split (subregion 3); Split (subregion 4); }
Recursive Merging If adjacent regions are Weakly split Weak edge Similar Similar greyscale/colour properties Merge them
Edge Following Detection Finds candidate edge pixels Following Links candidates to form boundaries
4/8 Connectivity Problem
Contour Tracking Scan image to find first edge point Track along edge points Spurs? Endpoints? Join edge segments
Edge Linking Aggregate collinear chain codes Colinear? Sequential least squares tolerance band
Sequential Least Squares Accumulate best fitting line to segments and error When error exceeds a threshold, finish segment Tolerance Band Accumulate best fitting line to segments If new point lies more than from line, finish segment
Hop Along Algorithm
Examples An example would show an edge detected image There would be a record of the edge points constituting each edge segment
Scale Based Methods Structures observed depend on scale of observation
Analysis Processing of an image should be at a level of detail appropriate to structures being sought Image pyramid Wavelet transform
Image Pyramid Reducing resolution Pixels in each layer computed by averaging groups of pixels in layer below. Or Use scale dependent operators – e.g. Marr Hildreth.
Wavelet Transform Transform image data Select coefficients Reverse transform
Watersheds of Gradient Magnitude Compare geographical watersheds Divide landscape into catchment basins Edges correspond to watersheds
Algorithm Locate local minima Flood image from these points When two floods meet Identify a watershed pixel Build a dam Continue flooding
Example watersheds local minima
watershed point
dam
Representing Regions Constituent pixels Boundary pixels
Region map As an array of region labels Pixel value = region label
Summary Region detection Growing Edge following Watersheds
I think there is a world market for maybe five computers Thomas J Watson, chairman IBM, 1943