Study On Algorithms Of Image Segmentation

Image segmentation is a process of separating an image into several un-lapped regions, in each which all pixels have similarity in intensity, color, texture, ect. Image segmentation, as the basis of image analysis and understanding, is one of the most difficult problems in computer vision. Taking the gray, color and texture images as examples and research methods of image segmentation along with their main applications are surveyed in details in this paper. The existing image segmentation algorithms are analyzed and studied deeply and the modified algorithms corresponding to the existing segmentation algorithms are proposed, mainly including: Firstly, aiming at the information loss in the process of the image segmentation, the concept of the "entropy" is introduced. With respect to the new definition of the image "entropy", researches are performed on the image segmentation based on the "entropy". Then, comparisons among these algorithms are done. According to the results, the limitations of these algorithms and the appropriate application situations in which these algorithms can work well are demonstrated in this paper. Secondly, by combining the hue information with the spatial information, a new method of color image segmentation based on the fusion of hue histogram and the spatial information is put forward. The experimental results demonstrate that the color image segmentation results of the proposed algorithm hold favorable consistency in terms of human perception. Thirdly, a texture segmentation method based on the Gabor and PCA is put forward after analyzing the current research methods of the texture image segmentation. Compared to the original Gabor Wavelet algorithm, the Gabor and PCA algorithm is less computational complexity. Fourthly, simulation researches are performed on the proposed segmentation algorithms. Finally, the summarization of this paper is given, and the future research is point out clearly.