Liver segmentation method of abdomen CT image, and CT imaging method thereof

Abstract

The invention discloses a liver segmentation method for an abdominal CT image. The method comprises the following steps: acquiring original abdominal CT image data and liver segmentation data; constructing a model training set and a model test set; constructing a liver segmentation original model; constructing a training loss function; training and testing the original liver segmentation model based on the training loss function by adopting the model training set and the model testing set to obtain a liver segmentation model; and performing liver segmentation on the target abdomen CT image by adopting the liver segmentation model, and performing morphological optimization to obtain a final liver segmentation result of the target abdomen CT image. The invention further discloses a CT imaging method comprising the liver segmentation method of the abdomen CT image. The continuity between CT sequence liver segmentation result slices can be improved, end-to-end accurate segmentation of the liver in the abdomen CT image is realized, the boundary precision is improved, and the robustness, the segmentation effect and the reliability are good.

Description

technical field [0001] The invention belongs to the field of image processing, and in particular relates to a liver segmentation method of an abdominal CT image and a CT imaging method thereof. Background technique [0002] With the development of economy and technology and the improvement of people's living standards, people pay more and more attention to health. [0003] The liver is one of the most important organs in the human body. Abdominal CT images are currently one of the most important means for medical personnel to obtain the status of the liver. Therefore, it is particularly important to perform complete and accurate liver segmentation from abdominal CT images. [0004] However, the liver has the characteristics of low contrast with adjacent organs, weak borders, and large differences in liver morphology between cases. It is still difficult to accurately segment the liver in abdominal CT images. At the same time, due to the large number of slices in CT images,...

AI Technical Summary

Problems solved by technology

In terms of neural network structure design, the currently commonly used two-dimensional convolutional neural network can only obtain the convolution features in the CT image slice, completely ignoring the context information between the CT sequence slices, while the three-dimensional convolutional neural network uses the three-dimensional convolution kernel Convolutional features containing contextual information between slices can be obtained, but the number of learnable parameters is too large and requires a lot of computing resources

In addition, in terms of loss function design, the current common technology is only based on the traditional loss function, that is, the area loss, but the current method has the disadvantage of not being sensitive enough to boundary errors, resulting in low segmentation boundary accuracy

Images