A brain tissue extraction method based on total convolution neural network

Abstract

The invention discloses a brain tissue extraction method based on a full convolution neural network, which comprises the following steps: firstly, a full convolution integral cutting network is used for preliminary segmentation of a two-dimensional original nuclear magnetic resonance image to obtain a preliminary segmentation result; secondly, a full convolution integral cutting network is used for preliminary segmentation of the two-dimensional original nuclear magnetic resonance image. Secondly, according to the preliminary segmentation results, the internal and boundary information of braintissue is separated. Thirdly, these pixels which can not be determined as brain tissue are selected as boundary candidate pixels, and these candidate pixels and their neighborhoods are sent to convolution neural network for secondary segmentation to realize classification and judgment. Finally, we integrate the internal segmentation results and the boundary segmentation results obtained from thesecondary segmentation, and obtain the final brain tissue extraction segmentation results. The invention carries out thick and thin twice segmentation, ensures the calculation efficiency of the method, realizes the fine segmentation goal, can be better applied to the brain magnetic resonance image, and realizes more accurate separation of brain tissue from non-brain tissues such as skull, eyeball,skin, fat and the like.

Description

technical field [0001] The invention belongs to the technical field of digital image processing, and relates to a brain tissue extraction method based on a fully convolutional neural network. Background technique [0002] The brain is one of the vital organs of the human body and an important part of our body. Human beings have never stopped studying the brain. Scientists hope to explore the unknown functions of the human brain by studying the complex structure inside the brain. Magnetic Resonance Imaging (MRI) technology is non-invasive, contains a large amount of information, and has the characteristics of multi-directional imaging. In MRI images, soft tissues with relatively low gray scale can be clearly distinguished. Therefore, important information such as the location and size of brain tissue anatomy in MRI images can be identified with the naked eye. In addition, MRI images have been widely used clinically because of their high signal-to-noise ratio, high resoluti...

AI Technical Summary

Problems solved by technology

At present, the existing traditional brain tissue extraction methods are difficult to achieve accurate segmentation of brain tissue

Images