Evolution Analysis of Adaptive Distance Preserving Level Set in Image Segmentation Based on Deep Learning
Download as PDF
As we know, image segmentation is the key technology from image processing to image analysis, and variational level set method has become one of the main development directions in the field of image segmentation. Adaptive distance preserving level set evolution model is one of the common methods. Although this method gets rid of the dependence of evolution curve on the initial position, it still has low segmentation accuracy and high accuracy For this reason, this paper analyzes the evolution of adaptive distance preserving level set in image segmentation based on deep learning. In this paper, convolution neural network is introduced into the evolution of adaptive distance preserving level set. Based on the adaptive distance preserving level set evolution model, a new variable weight coefficient is introduced, and a new edge stop function is defined. Experiments on different data sets are carried out to verify the reliability of the proposed method. The results show that in the traditional adaptive distance preserving level set evolution model, the average value of signal-to-noise ratio is 22.1343, and the average value of peak signal-to-noise ratio is 28.0425. In this method, the average value of signal-to-noise ratio is 24.2743, and the average value of peak signal-to-noise ratio is 29.9982. Therefore, the denoising effect of this method is better. In addition, the experimental results show that the segmentation accuracy of the proposed method is much higher than that of the traditional adaptive distance preserving level set evolution model. Therefore, this method can effectively improve the segmentation accuracy and denoising effect.
Deep Learning, Image Segmentation, Convolution Neural Network, Level Set Evolution