Research Of Image Segmentation Based On Improved CV Model

Image segmentation is an important branch of image processing technology and is atraditional and challenging project. Image processing relies on the quality of image segmentation,and which directly affects then results of subsequent image analysis, interpretation andidentification. Therefore, image segmentation in the practical application and theoretical study ofthe field has a wide attention.Although PDE (Partial Differential Equations) method involves fairly large calculationscompared with other image segmentation methods, this flexible numerical method has a goodstability when used in disserting, and the PDE can meet the precision of image segmentation, andwith high quality when used with image restoration, etc. Thus, improve the accuracy and speedof the method has become a research hotspot in related field.In this article, we first introduces the purpose and significance of image segmentation; themain segmentation method of digital image segmentation, and introduces the basic principle,application of active contour model, level set method. After that, based on the CV(Chan-Vese)model and analysis of the improved algorithm of CV model, we propose two new segmentationmethods:(1) Adaptive CV model combined with region gradient: Based on the regional gradientalgorithm, introduces the adaptive weight and uses the weighted average to calculate thefitting center. Experimental results show this new model has better robustness to the location ofinitialization.(2)Image segmentation based on Otsu and improved CV model: Based on theanalysis of Otsu algorithm and CV model, combined the Otsu algorithm with improved CVmodel, we presents a new segmentation method; Experiments and results analyzing indicate thatthe new proposed method is effective. Finally, main work and the direction of further researchare summarized in this paper.