Multi-threshold image segmentation based on an improved differential evolution

data mining

Image processing plays an important role in computer vision. The process of image segmentation provides the partition of image into different segments according to their feature attribute. Region based segmentation is a type similarity based segmentation. Another type of segmentation is called thresholding based segmentation. In thresholding based segmentation method some thresholding techniques are used. Thresholding techniques are classified into two major categories as, Global and Local. In global thresholding, pixel values are categorized in two classes, one class belong to object and another class belong to background. We use one threshold value in global thresholding for whole image that belongs to single level thresholding and if threshold value used in segmentation is more than one, technique is called multilevel thresholding. Local thresholding belongs to multilevel thresholding method. In this paper a comparative analysis of global thresholding and local thresholding methods is made according to time taken for image segmentation. Experimental results provide a conclusion that Global thresholding takes less time than local thresholding.

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.