Downloaded 1,791 times
Uploaded byMostafa G. M. Mostafa
Digital Image Processing: Image Segmentation
This document discusses image segmentation techniques. It describes how segmentation partitions an image into meaningful regions based on discontinuities or similarities in pixel intensity. The key methods covered are thresholding, edge detection using gradient and Laplacian operators, and the Hough transform for global line detection. Adaptive thresholding is also introduced as a technique to handle uneven illumination.
Digital Image Processing: Image Segmentation
- 1. CSC447: Digital Image Processing Chapter10: Prof. Dr. Mostafa Gadal-Haqq M. Mostafa Computer Science Department Faculty of Computer & Information Sciences AIN SHAMS UNIVERSITY
- 2. Segmentation attemptsto partition the pixels of an image into groups that strongly correlate with the objects in an image Typically the first step in any automated computer vision application Image Segmentation 2CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 3. Image Segmentation • Segmentationalgorithms generally are based on one of two basis properties of intensity values • Discontinuity: to partition an image based on abrupt changes in intensity (such as edges) • Similarity: to partition an image into regions that are similar according to a set of predefined criteria. 3CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 4. Image Segmentation ImageSegmentation 4CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 5. Image Segmentation ImageSegmentation 5CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 6. Image Segmentation Detectionof discontinuities: There are three basic types of gray-level discontinuities: points , lines , edges the common way is to run a mask through the image 6CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 7. Point Detection: • Notethat the mark is the same as the mask of Laplacian Operation (in chapter 3) • The only differences that are considered of interest are those large enough (as determined by T) to be considered isolated points. |R| >T 7CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 8. Point Detection: 8CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 9. Line Detection • Horizontalmask will result with max response when a line passed through the middle row of the mask with a constant background. • the similar idea is used with other masks. • note: the preferred direction of each mask is weighted with a larger coefficient (i.e.,2) than other possible directions. R1 R2 R3 R4 9CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 10. Line Detection • Applyevery masks on the image • let R1, R2, R3, R4 denotes the response of the horizontal, +45 degree, vertical and -45 degree masks, respectively. • if, at a certain point in the image |Ri| > |Rj|, for all j≠i, • that point is said to be more likely associated with a line in the direction of mask i. 10CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 11. Line Detection • Alternatively,if we are interested in detecting all lines in an image in the direction defined by a given mask, we simply run the mask through the image and threshold the absolute value of the result. • The points that are left are the strongest responses, which, for lines one pixel thick, correspond closest to the direction defined by the mask. 11CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 12. Line Detection 12CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 13. Edge Detection Approach Segmentation by finding pixels on a region boundary. Edges found by looking at neighboring pixels. Region boundary formed by measuring gray value differences between neighboring pixels 13CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 14. Edge Detection • anedge is a set of connected pixels that lie on the boundary between two regions. • an edge is a “local” concept whereas a region boundary, owing to the way it is defined, is a more global idea. 14CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 15. Edge Detection 15CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 16. Edge Detection 16CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 17. Edge Detection 17CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 18. Edge Detection Detectionof discontinuities: Image Derivatives 18CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 19. Edge Detection • Firstcolumn: images and gray- level profiles of a ramp edge corrupted by random Gaussian noise of mean 0 and = 0.0, 0.1, 1.0 and 10.0, respectively. • Second column: first-derivative images and gray-level profiles. • Third column : second- derivative images and gray- level profiles. 19CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 20. Edge Detection GradientOperator )()( )()( mask33for 741963 321987 2/122 zzzzzzG zzzzzzG GGf y f x f G G f y x yx y x 20CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 21. Edge Detection Prewittand Sobel Operators 21CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 22. Edge Detection 22CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 23. Edge Detection 23CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 24. Edge Detection 24CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 25. Edge Detection 25CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 26. Edge Detection TheLaplacian 26CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 27. Edge Detection 27CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 28. Edge Detection TheLaplacian of Gaussian (LoG) 28CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 29. Edge Detection TheLaplacian of Gaussian (LoG) 29CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 30. The Hough Transform 30CSC447:Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 31. The Hough Transform Global processing: The Hough Transform 31CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 32. The Hough Transform Global processing: The Hough Transform 32CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 33. The Hough Transform Global processing: The Hough Transform 33CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 34. Region-Based Segmentation 34CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 35. What is aRegion? Basic definition :- A group of connected pixels with similar properties. Important in interpreting an image because they may correspond to objects in a scene. For that an image must be partitioned into regions that correspond to objects or parts of an object. 35CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 36. Region-Based vs. Edge-Based Region-Based Closed boundaries Multi-spectral images improve segmentation Computation based on similarity Edge-Based Boundaries formed not necessarily closed No significant improvement for multi-spectral images Computation based on difference 36CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 37. Image Thresholding •What isthresholding? •Simple thresholding •Adaptive thresholding 37CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 38. Thresholding – AKey Aspect Most algorithms involve establishing a threshold level of certain parameter. Correct thresholding leads to better segmentation. Using samples of image intensity available, appropriate threshold should be set automatically in a robust algorithm i.e. no hard-wiring of gray values 38CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 39. Automatic Thresholding Useof one or more of the following:- 1. Intensity characteristics of objects 2. Sizes of objects 3. Fractions of image occupied by objects 4. Number of different types of objects Size and probability of occurrence – most popular Intensity distributions estimate by histogram computation. 39CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 40. Automatic Thresholding Methods Some automatic thresholding schemes: 1. P-tile method 2. Iterative threshold selection 3. Adaptive thresholding 40CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 41. Thresholding Methods P-tileMethod:- If object occupies P% of image pixels then set a threshold T such that P% of pixels have intensity below T. Iterative Thresholding:- Successively refines an approx. threshold to get a new value which partitions the image better. 21 2 1 T 41 CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 42. P-Tile Thresholding • Thresholdingis usually the first step in any segmentation approach • Single value thresholding can be given mathematically as follows: Tyxf Tyxf yxg ),(if0 ),(if1 ),( 42CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 43. P-Tile Thresholding • Basicglobal thresholding: • Based on the histogram of an image Partition the image histogram using a single global threshold 43CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 44. P-Tile Thresholding • Basicglobal thresholding: • The success of this technique very strongly depends on how well the histogram can be partitioned 44CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 45. Iterative P-Tile Thresholding •The Basic global thresholding: 1. Select an initial estimate for T (typically the average grey level in the image) 2. Segment the image using T to produce two groups of pixels: G1 consisting of pixels with grey levels >T and G2 consisting pixels with grey levels ≤ T 3. Compute the average grey levels of pixels in G1 to give μ1 and G2 to give μ2 45CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 46. Iterative P-Tile Thresholding •The Basic global thresholding: 4. Compute a new threshold value: 5. Repeat steps 2 – 4 until the difference in T in successive iterations is less than a predefined limit T∞ This algorithm works very well for finding thresholds when the histogram is suitable. 2 21 T 46CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 47. P-Tile Thresholding 47CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 48. P-Tile Thresholding 48CSC447: DigitalImage Processing Prof. Dr. Mostafa GadalHaqq.
- 49. P-Tile Thresholding • Limitationof P-Tile thresholding: 49CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 50. P-Tile Thresholding • Limitationof P-Tile thresholding: 50CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 51. P-Tile Thresholding • Limitationof P-Tile thresholding: 51CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 52. Adaptive Thresholding AdaptiveThresholding is used in scenes with uneven illumination where same threshold value not usable throughout complete image. In such case, look at small regions in the image and obtain thresholds for individual sub-images. Final segmentation is the union of the regions of sub-images. 52CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 53. Adaptive Thresholding Thresholding– Basic Adaptive Thresholding 53CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 54. Adaptive Thresholding Thresholding– Basic Adaptive Thresholding 54CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 55. Adaptive Thresholding Thresholding– Basic Adaptive Thresholding 55CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 56. Adaptive Thresholding Thresholding– Basic Adaptive Thresholding 56CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 57. Adaptive Thresholding Thresholding– Basic Adaptive Thresholding 57CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.
- 58. Summary Segmentation isthe most essential step in most scene analysis and automatic pictorial pattern recognition problems. Choice of the technique depends on the peculiar characteristics of individual problem in hand. 58CSC447: Digital Image Processing Prof. Dr. Mostafa GadalHaqq.