Lung tissue segmentation method and device based on capsule network, equipment and storage medium

Abstract

The invention discloses a lung tissue segmentation method based on a capsule network, and the method comprises the steps: inputting a to-be-segmented chest CT image into a convolution layer of a segmentation network, and obtaining a first feature map; inputting the first feature map into a capsule layer of the segmentation network to obtain a second feature map; and finally, performing binarization processing on the second feature map according to the modulus value of the multi-dimensional vector corresponding to each pixel point of the second feature map, so that a lung tissue segmentation image can be obtained. Therefore, the bottom-layer features of the chest CT image are extracted through the convolution layer of the segmentation network, the bottom-layer features are subjected to vectorization operation through the capsule layer of the segmentation network, features with higher dimension and a hierarchical structure can be learned in the training and learning process, and therefore when lung tissue in the chest CT image is segmented, the difference between targets can be described more accurately, and accurate lung tissue segmentation of the chest CT image is achieved.

Description

technical field [0001] The present invention relates to the technical field of medical image processing, in particular to a capsule network-based lung tissue segmentation method, device, equipment and storage medium. Background technique [0002] Medical image segmentation is one of the important research directions in the intersection of computer vision and medicine. Especially in recent years, the image segmentation method based on deep learning has developed extremely rapidly in the field of image semantic segmentation, and has made great achievements. A large number of excellent network structures have emerged. , such as Unet, fully convolutional neural network FCN, deep convolutional autoencoder SegNet and dense neural network DenseNet, etc. At present, the research results of lung segmentation using the above model have been published in mainstream academic journals. [0003] When applying the above-mentioned deep learning method for semantic segmentation of chest CT ...

AI Technical Summary

Problems solved by technology

like figure 1 As shown, the two image slices are taken from the same chest CT image, and are manually calibrated, and some slices ( figure 1 a) The trachea was excluded when segmenting lung tissue, others ( figure 1 b) The trachea is included in the lung tissue, so when the above deep learning method is directly used for lung segmentation, it is difficult to obtain a large number of accurately calibrated training sets, so the segmentation accuracy obtained is limited

Images