Image Segmentation In The Footprint Segmentation

Image segmentation is an extremely important one in the field of image processing, it is measured at some image’s distinguishing feature, making the images divided into some regions that have Some Certain Significance.The traditional image segmentation method involves some other methods which based on Image edge segmentation or Threshold segmentation. In recent years, new technologies and new theories are widely applied to image segmentation. Many researchers have proposed large numbers of newly image segmentation methods, such as Image segmentation methods which based on watershed, spectral clustering and interactive theoretic. Although, so far researchers have made thousands of image segmentation algorithm, but a general theory of segmentation has not been founded. That is to say, now mostly proposed segmentation algorithms are issue-specific, not suitable for all image segmentation algorithms.Footprint recognition has been long and widely used in the field of criminal investigation field and pathology, but the footprint image analysis must footprint image extracted, and the footprint image between the toes and the soles of the feet, toes and between the toes separated, and the segmentation results of a direct impact on the footprint of the image analysis. Some of the current segmentation algorithm can not be good to meet the Footprint Segmentation requirement.Automatic segmentation of footprint image, this paper presents the algorithm based on spectral clustering footprint image segmentation variable. First of all, through the acquisition of image denoising, edge detection, image binarization footprint image is extracted, to form a binary image. Then, extract the edge of the footprint image of the concave point detection, and spectral clustering based on the footprint of the characteristics of the image will be extracted from the concave point. Finally, according to the concave point of the cluster matches, to complete the footprint image segmentation.The paper presents experimental results of the footprint image segmentation, that this method can achieve automatic segmentation of the footprint image and segmentation results more desirable.