CodeForge.comFree Open Source Codes Forge and Sharing |
MATLAB image segmentation program
Iterative methods
Iterative-threshold of the basic idea is : First of all, according to the distribution of objects in an image to grayscale, select an approximate threshold as the initial threshold, a better approach is to image gray value as the initial threshold; and then by dividing the image and modify the iteration process of threshold values recognized the best threshold [5] 。 Iterative-threshold selection process can be described as follows.
(1) Select an initial threshold T 。
(2) Using threshold values T Split into two sets of image given image, remember to R 1 ,R 2 。
(3) calculate R 1 and R 2 's average values u1 and u2 。
(4) Select the new threshold value T , and
T= ( u 1+ u 2 ) /2 (2-4 )
(5) Repeat (2)~(4) until R 1 and R 2 's average values u 1 and u 2 no longer changes.
When implemented, first under the initial switching function will enter the figure images one by one into the foreground and background, after the end of the first pass on the scanned images, the average values of the two integrals to determine a threshold. The threshold control switch will once again enter into the foreground and background, and do a new switch function. Repeated iterations until the switching function is not changed, get the foreground and background is the final segment results. Iterative thresholding of images derived from to good effect. Thresholds based on iteration can distinguish between the foreground and background of the main area of the image, but don't have very good images of subtle distinction. For certain images, micro-data changes will cause great changes in Iris, both of data is only slightly changed, but Iris has extreme contrast. Histogram peaks visible, bottom deep image, iterative methods can quickly get satisfactory results, but not obvious for the histogram peaks, or differences in image and background, and iterative method to the selected threshold than other methods.
| Name | Size | Date |
| 41.54 kB | 16-07-01 23:24 | |
| 0.00 B | 30% 06-05-12 | |
| 121.30 kB | 06-05-12 21:46 | |
| 177.83 kB | 06-05-12 21:40 | |
| 146.33 kB | 27-07-01 22:52 | |
| 136.00 B | 06-05-12 21:51 | |
| 131.00 B | 15-05-12 08:27 | |
| 674.00 B | 15-05-12 08:27 | |
| 296.00 B | 15-05-12 08:28 | |
| 318.00 B | 15-05-12 08:29 | |
| 12.64 kB | 15-05-12 08:38 | |
| 0.00 B | 15-05-12 08:39 |