Infant brain magnetic resonance image partitioning method based on fully convolutional network

Abstract

The invention provides an infant brain magnetic resonance image partitioning method based on a fully convolutional network. The main content of the method comprises the multi-stream three-dimensionalfully convolutional network (FCN) with jump connection, partial transfer learning, training and testing and evaluation. According to the process of the method, first, a probability graph of each pieceof brain tissue is learned from a multi-modal magnetic resonance image; second, initial partitions of different brain tissue are obtained from the probability graphs and used for calculating a distance graph of each piece of brain tissue; third, spatial context information is simulated according to the distance graphs; and last, final partitioning is realized by use of spatial correlation information and the multi-modal magnetic resonance image, wherein the training process mainly comprises training data increasing, training patch preparation and iterative training, and various detection values are used for evaluation after testing is performed. Through the partitioning method, a white matter area, a grey matter area and a cerebrospinal fluid area are successfully divided, the potential gradient vanishing problem of multi-level deep supervision is relieved, training efficiency is improved, and partitioning performance is greatly enhanced.

Description

technical field [0001] The invention relates to the field of image segmentation, in particular to a method for segmenting infant brain magnetic resonance images based on a fully convolutional network. Background technique [0002] Research on region of interest segmentation of medical images is the most important basis in medical image analysis. Accurate, robust and fast image segmentation is the most important step before quantitative analysis, 3D visualization and other follow-up links. Important clinical applications such as radiotherapy planning and treatment evaluation have laid the most fundamental foundation. Using automatic, accurate and quantitative computer-aided image analysis can help clinicians and researchers process massive image information efficiently and accurately. Brain segmentation in magnetic resonance (MR) images is a central part of this quantitative analysis tool, as it provides quantitative measurements of different brain structures and provides ba...

AI Technical Summary

Problems solved by technology

[0004] Aiming at the problem that gray matter and white matter have almost the same intensity level, the purpose of the present invention is to provide a method for brain MRI segmentation of infants and young children based on a fully convolutional network. Probability maps of the tissues, and then an initial segmentation of different brain tissues is obtained from the probability maps, which are used to calculate distance maps for each brain tissue, followed by modeling spatial context information from the distance maps, and finally by using spatial correlation information and multimodal magnetic resonance images To achieve the final segmentation, the training process mainly includes increasing training data, preparing training patches and iterative training, and evaluating with various detection values ​​after testing

Images