Active Contour Image Segmentation with Mixed Gaussian Distribution Model

Abstract

The invention discloses an active contour image segmentation method for a Gaussian mixture model. The method comprises steps of according to the conventional local Gaussian probability model LGDF, providing a new energy function which considers Gaussian distribution of the gray scale in the global range and the Gaussian distribution of the gray scale in the local range of the image at the same time and which integrates the Gaussian distribution models of two different ranges, so as to effectively utilizes functions of global information and local information of the image in image segmentation; in addition, in order to prevent the curve length problem caused by smoothness and unreasonable bending of the segmentation result curve, introducing two different punishment bound terms of distance displacement and curve length into the energy function; and finally, solving the energy function by a variational level set method to achieve automatic extraction of the object contour. The active contour image segmentation method for the Gaussian mixture model can accurately extract the target object in the image in conditions of different image backgrounds and grayscale uniformities.

Description

technical field [0001] The invention belongs to the field of image segmentation and target detection, and in particular relates to an active contour image segmentation method of a mixed Gaussian distribution model. Background technique [0002] Image segmentation is an image processing technology that separates the target object and the background area in the image according to the grayscale characteristics of the image. This technology can deepen people's understanding of image understanding and analysis, computer vision, and target detection. research and wide application. In the field of image segmentation, various algorithms have been proposed to propose target objects, including region growing method, classification-based segmentation method, watershed algorithm, and active contour model algorithm; among them, the segmentation algorithm based on active contour model is relatively A popular research branch, this type of algorithm can obtain segmentation results with hig...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the above-mentioned technical problems, the present invention mainly aims at the above-mentioned deficiencies in the region-based active contour model segmentation algorithm, and proposes an active contour image segmentation method based on a mixed Gaussian distribution model to solve complex background information and serious grayscale inhomogeneity in the image The segmentation results caused by the characteristics are rough, the accuracy is low, and it is time-consuming, so as to accurately and completely extract the target objects with different shapes and sizes in the image

Images