An Image Segmentation Method Based on pca Reconstruction Error Level Set

Abstract

The invention belongs to the technical field of image processing, and discloses an image segmentation method based on a PCA reconstruction error level set, which includes inputting an image to be segmented; extracting image features; initializing an evolution curve; using PCA to obtain basis vectors of internal and external areas of the image; According to the basis vector, reconstruct each pixel in the image; calculate the reconstruction error of each pixel in the image; accumulate the reconstruction error of each pixel, construct the data-driven energy item; minimize the novel energy function, and drive the curve Evolution, get the segmentation result. With respect to the image segmentation of the level set of the segmented constant model, the PCA technology used in the present invention assumes that the image information satisfies the Gaussian distribution, so it can better segment the inhomogeneous image and is robust to noise; compared to the image of the Gaussian model level set For segmentation, the calculation of the Gaussian model is time-consuming, but the PCA technique used in the present invention runs faster.

Description

technical field [0001] The invention belongs to the technical field of image processing, in particular to an image segmentation method based on PCA reconstruction error level set. 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 active contour model is a kind of segmentation method which is very popular in current research. Fundamentally speaking, the active contour model method 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; the geometric type is to embed the low-dimensional curve into a higher one-dimensional surface. In , the surface is called a level set function, and the original curve is its zero level set, and is describe...

AI Technical Summary

Problems solved by technology

However, the parameters of the Gaussian model are time-consuming to solve, resulting in a slow speed of this type of method.

[0006] In summary, the problems in the prior art are: 1) The segmentation of non-homogeneous images is not accurate enough, and it is not robust enough to noise, for example, the image segmentation based on the level set of the segmental constant model; 2) The operation speed is slow, for example, the image segmentation based on the level set of the Gaussian model

[0007] The difficulty and significance of solving the above technical problems: The difficulty in solving the above problems is not only to overcome the shortcomings of inaccurate segmentation of inhomogeneous images and not robust to noise, but also to overcome the shortcomings of slow running speed

Images