The Research Of Image Segmentation Based On Fractal And Active Contour Model

Image segmentation method is used to extracting the region of interest, which plays an important research project significance. For decades, foreign and domestic scholars and researchers have researched on image segmentation in depth and developed many new technologies, one of them is the active contour model.We gave an improved methods based on researching the applications of the active contour model in the field of image segmentation and summing up the advantages and disadvantages of existing methods. For the dependence of the initial contour position in traditional parameters active contour model, as well as the difficulties progressing into boundary concavities, the paper presents a threshold of fractal dimension method to determine the initial contour inspired by the fractal dimension of image edge detection. First of all, calculate fractal dimension of each region in the original image, for the different target regions of interest select the appropriate threshold, identify the pixels that meet the threshold of the fractal dimension. Then, using B-spline curve fitting of these pixels as the initial contour position. Finally, the ultimate aim to achieve a correct segmentation is use of GVF Snakes model for target-contour approaches. Experimental results show that the region of interest extracted through this method can realize self-adaptive image segmentation, not only the score-shaped edge detection more accurate, but also reduce the number of iterations of GVF snakes model and the difficulties of parameter adjustment. It makes good segmentation results for low SNR medical image.