Snake Model Improvement Based on Bilateral Filter

Abstract

In view of the limitation of the extraction of the target object's edge contour by the classic Snake model, an improved algorithm of Snake model based on bilateral filter was proposed in this paper. This algorithm uses a bilateral filter instead of the Gaussian filter in the classic Snake model to preprocess the target image. Through comprehensive consideration of the spatial domain of the target image noise point and the gray value of adjacent pixels, the target image is more accurate and effective pretreatment, and then the target image obtained at this time after the pre-processing is finished, the subsequent edge contour extraction processing and operation are performed again, and the Matlab R2015b experimental platform and its own standard experimental image are used for experimental simulation. The experimental results show that the improved model in this paper is significantly better than classic Snake model in extracting the edge contour of the target object, which can not only better protect the edge of the target image, but also have a good optimization effect on the overall definition of the target image. During the experiment, the improved Snake model was used to observe and analyze the extraction effect of the edge contour of the same target object to verify the accuracy and effectiveness of the improved model proposed in this paper.