Image segmentation method based on simplified local binary fitting (LBF) model

Abstract

The invention discloses an active contour model image segmentation method based on non-subsampled contourlet transformation. The active contour model image segmentation method comprises the steps that firstly, an image to be segmented is expressed with multiresolution through the non-subsampled contourlet transformation, secondly, a probability model with a multiresolution coefficient is established, and finally, integration operations of the multiresolution coefficient are carried out by utilizing the active contour model based on regions, so that the aim of image segmentation is achieved. The experimental results show that the active contour model image segmentation method can nicely carry out image segmentation operations, integration of the segmented image is ensured, and detail information of the image can be segmented.

Description

technical field [0001] The invention belongs to the technical scope of image processing, and in particular relates to an image segmentation method based on a simplified LBF model that can improve image segmentation accuracy and segmentation efficiency. Background technique [0002] Image segmentation is one of the important tasks in image processing, and its purpose is to separate the object of interest in the image from the rest of the image, so as to serve higher-level image processing. There are currently two main models: the segmentation model based on edge information and the segmentation model based on region information. Edge-based segmentation models cannot accurately segment images with blurred or edgeless edges, while C-V models based on global region segmentation overcome the shortcomings of edge models. The region segmentation model uses the information of the inner and outer regions of the active contour to eliminate the interference of some noises, can deal wi...

AI Technical Summary

Problems solved by technology

Among them, the LBF model is a segmentation model based on local area information, which can handle the texture details of the image better than the C-V model, and the processing effect is better, but due to the processing of local information, its segmentation efficiency is low, and for the initial contour line The selection is not very flexible, and there are many artificial factors

There are three deficiencies in the existing image segmentation methods based on partial differential equations: first, the existing model simply uses the global information of the image (simple background and target), and the processing effect is relatively rough; second, the LBF model The design of some items (such as length items, etc.) is more complicated and cumbersome, which affects the convergence speed of the model; third, the evolution of the level set function of the C-V model needs to be initialized repeatedly, while the LBF model needs to reasonably design the position of the initial contour line. In practical application, it will be greatly restricted, and the implementation will be more time-consuming

Images