Independent component analysis multi-shape prior level set method, image segmentation system

Abstract

The present invention belongs to the field of image processing technology, and a multi -shape acetral level method and image segmentation system are released.Code of the horizontal set function of the priority; the formation of the formation of the shape; the independent component analysis is used to reduce the dimension; projected the current horizontal set function to the low -dimensional space;The combination of items constitutes the energy function; minimizes the energy function, the driving curve evolution, and the segmentation results are obtained.The present invention can eliminate the characteristics of high -dimensional redundancy in shape, and then more accurate statistical distribution of shape priority, forming more effective shape constraints, and finally more accurate segmentation results.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an independent component analysis multi-shape prior level set method and an image segmentation system. Background technique [0002] In the past half century, image segmentation has attracted people's attention and maintained a persistent research enthusiasm. So far, thousands of segmentation methods based on different theories have been proposed. Among them, the method based on the active contour model is a kind of segmentation method that is currently researched very popular because it uses prior knowledge to guide the segmentation process and provides a unified framework for segmentation. Fundamentally speaking, active contour modeling methods can be divided into two types: parametric and geometric. The parametric refers to the curve or surface that directly expresses the deformation in the form of parameters; In three-dimensional surfaces, as its zero level set, and...

AI Technical Summary

Problems solved by technology

[0003] To sum up, the problem existing in the existing technology is: in the current method of introducing shape prior, for the single prior method, there is a single shape prior that cannot reflect the commonality of a class of shapes, when the difference between the target to be segmented and the shape prior When it is relatively large, the problem that better results cannot be obtained; for the multi-shape prior method, the prior shape is generally sparsely distributed in the high-dimensional feature space, it is difficult to accurately estimate the distribution of the shape prior, and the inaccurate distribution estimation Issues that can lead to inaccurate segmentation results

Images