digital image processing-8
TRANSCRIPT
-
8/6/2019 Digital Image Processing-8
1/38
-
8/6/2019 Digital Image Processing-8
2/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation Watershed Segmentation
RepresentationIntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
3/38
Introduction Image segmentation is the process of partitioning the
digital image into multiple regions that can be
associated with the properties of one or more objects
-
8/6/2019 Digital Image Processing-8
4/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation Watershed Segmentation
RepresentationIntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
5/38
TextureTexture provides measures of properties such assmoothness, coarseness, and regularity.
-
8/6/2019 Digital Image Processing-8
6/38
-
8/6/2019 Digital Image Processing-8
7/38
Co-Occurance MatrixLet P be a position operator, and A a k x k matrix.aij shows the number of times that pixels with gray
level zi occur at position given by P relative to points with gray level z j.
Matrix A is called co-occurance matrix and can provide
statistical properties of the texture.
-
8/6/2019 Digital Image Processing-8
8/38
Example Assume P is one pixel to the right and one pixel belowGray level values are : 0, 1, and 2
Image data:
Co-occurance matrix is:
-
8/6/2019 Digital Image Processing-8
9/38
Statistical Moments of TextureLet Matrix C be formed by dividing every element of A by the numberof point pairs that satisfy P.The following moments are defined to compare textures:
-
8/6/2019 Digital Image Processing-8
10/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
11/38
Level Set SegmentationInstead of manipulating the contour directly, thecontour is embedded as the zero level set of a higherdimensional function called the level-set function y( X, t).The level-set function is evolved under the control of adifferential equation.
At any time, the evolving contour can be obtained by extracting the zero level-set G(( X), t) = {y(X, t) = 0}from the output
-
8/6/2019 Digital Image Processing-8
12/38
Zero Set in a Level Set
-
8/6/2019 Digital Image Processing-8
13/38
Example
-
8/6/2019 Digital Image Processing-8
14/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
15/38
Watershed Segmentation Algorithm:
Convert the gray level image into a topographic image where the height of each point is proportional to its gray level intensity.Punch a hole at each region minimum at let the wholetopography be flooded from below.The points where the water from different regions joinare boundaries of the regions
-
8/6/2019 Digital Image Processing-8
16/38
Watershed Segmentation
-
8/6/2019 Digital Image Processing-8
17/38
Example
-
8/6/2019 Digital Image Processing-8
18/38
Example
-
8/6/2019 Digital Image Processing-8
19/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
20/38
RepresentationThe result of segmentation should be represented anddescribed in a form suitable for further computerprocessing.
A region can be represented in terms of its externalcharacteristics (boundary). A region can be represented in terms of its internalcharacteristics.
-
8/6/2019 Digital Image Processing-8
21/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
22/38
Chain CodesChain codes are generated by following a boundary in aclockwise or counter-clockwise direction and assigning adirection to the segments connecting every pair of pixels.
Disadvantage: Can be unacceptably long.Solution: Re-sampling (down sample) the boundary
Disadvantage: Is starting point dependentSolution: Normalize the representation string to thesmallest integer.
-
8/6/2019 Digital Image Processing-8
23/38
Chain Code Directions
-
8/6/2019 Digital Image Processing-8
24/38
Sample Chain Code
-
8/6/2019 Digital Image Processing-8
25/38
-
8/6/2019 Digital Image Processing-8
26/38
-
8/6/2019 Digital Image Processing-8
27/38
Polygonal Approximation A boundary can be represented with arbitrary accuracy by a polygon.The approximation is exact when the number of sidesis equal to the number of points in the boundary.Finding a polygonal representation can be very time-consuming.
-
8/6/2019 Digital Image Processing-8
28/38
-
8/6/2019 Digital Image Processing-8
29/38
Splitting Techniques
-
8/6/2019 Digital Image Processing-8
30/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
31/38
Signature A signature is a 1D representation of a boundary.e.g. Plotting distance to centroid as a function of angle
Invariant to translation
Disadvantages:Rotation and scaling dependant
Defined only for convex regions
-
8/6/2019 Digital Image Processing-8
32/38
Signature Example
-
8/6/2019 Digital Image Processing-8
33/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
34/38
Boundary SegmentsDecomposing a boundary into segments simplifiesrepresentation.Convex Hull can be used for decomposition.
A new segment can be started whenever a Convex Hulldeficiency is entered or exited.
-
8/6/2019 Digital Image Processing-8
35/38
Boundary Segments Example
-
8/6/2019 Digital Image Processing-8
36/38
TopicsSegmentationTexture Based SegmentationLevel Set Segmentation
Watershed SegmentationRepresentation
IntroductionChain Codes
Polygonal ApproximationsSignaturesBoundary SegmentsSkeletons
-
8/6/2019 Digital Image Processing-8
37/38
SkeletonThe structural shape of a region can be represented by a graph.The structural graph is obtained by thinning theregion and finding the skeleton.
-
8/6/2019 Digital Image Processing-8
38/38
Questions?