Image Segmentation and Denoising Algorithm Based on Partial Differential Equations

Abstract

The image denoising algorithm based on partial differential equation can selectively solve the problem of noise and image smoothing, which has become the focus of current image denoising algorithms. In the research of image segmentation, this paper proposes a new image segmentation model, which not only can segment multi-phase images, but also achieves simple implementation, which greatly saves the time required for segmentation. The total variation denoising model is analyzed to improve the fidelity of the model: (1) The total variation and the image denoising model, which constructs a trend fidelity term that can effectively suppress the “staircase effect”; 2) Wavelet transform image denoising model based on curvature variation, which uses a horizontal set of enhanced images to establish a curvature drive function based on the horizontal set curvature, and then introduces the curvature drive function as a correction factor into the variational model to establish the curvature change. The experimental results show that the two improved Partial differential equation image segmentation models (PDEISM) have obvious denoising effects, good visual effects and strong comprehensive performance. The simulation experiment was carried out with MATLAB software to verify the superiority of the improved algorithm. The conclusions drawn from the final simulation experiment are consistent with the conclusions of the theoretical analysis, which indicates that the improved algorithm is superior to the traditional algorithm in overall performance and has better practical application value.