Medical image segmentation method based on coding and decoding structure in combination with residual module

Abstract

The invention discloses a medical image segmentation method based on a coding and decoding structure in combination with a residual error module. A specific combination mode of an encoding and decoding network and a residual module is adopted for a convolutional neural network for medical image segmentation, and the encoding and decoding network can extract effective features without a large amount of data, so that the convolutional neural network can fully learn the features of an image; the residual error module solves the problem of network degradation possibly occurring in the network, andimproves the segmentation precision of the medical image. According to the method, a deep learning method is adopted, so that the convolutional neural network can automatically learn the features ofthe region needing to be segmented in the medical image through the data set, and a tedious operation process and parameter adjustment are not needed. In addition, the convergence rate of the networkis improved by using pre-training weight parameters and an edge optimization strategy based on a full-connection conditional random field, and a medical image segmentation result is further optimized.

Description

technical field [0001] The invention relates to the technical field of medical image segmentation, in particular to a medical image segmentation method based on an encoding and decoding structure combined with a residual module. Background technique [0002] Image segmentation has always been a hot research topic in the field of image processing and analysis, and is a crucial problem in computer vision. In medical image analysis, image segmentation technology plays an increasingly important role. Image segmentation is an indispensable means to accurately extract specific tissues or regions in the image. Segmented images are used for quantitative analysis and research on diseased areas, and can play an important auxiliary role in the diagnosis of doctors. At present, there are mainly the following methods for medical image segmentation: [0003] The first is the traditional image segmentation, which usually divides the image into different regions according to the grayscale...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the problems of low segmentation accuracy and segmentation efficiency and limited use of existing medical image segmentation methods, the present invention provides a medical image segmentation method based on an encoding and decoding structure combined with a residual module

Images