New feature extraction and segmentation method for liver CT image

Abstract

The invention discloses a new feature extraction and segmentation method for a liver CT image. The new feature extraction and segmentation method comprises the steps of obtaining liver tumor CT imagedata and performing resampling; dividing a region of interest of the image data by adopting a fuzzy clustering method to obtain the region of interest as a centroid and obtain a core learning sample;preprocessing the core learning sample; adopting a reinforcement learning search baseline network to form an adaptive three-dimensional convolution deep learning network; training the adaptive three-dimensional convolutional deep learning network by adopting the sample data to obtain a feature extraction and segmentation model; and carrying out feature extraction and segmentation on the liver CT image by adopting a feature extraction and segmentation model. According to the method, manual calibration conditions are not needed, higher-level abstract features can be extracted, feature segmentation of the liver image can be automatically and efficiently carried out, and the feature segmentation efficiency is higher; and meanwhile, the reliability is high, and the practicability is good.

Description

technical field [0001] The invention specifically relates to a new method for feature extraction and segmentation of liver CT images. Background technique [0002] With the rapid development of society, economy, education, and science and technology, people pay more and more attention to their own health. Scientific inspection and disease prevention are important measures to improve the quality of life and health. The CT image of the liver includes the pixel value, texture shape, texture distribution area, direction feature, geometric feature, edge definition and other features of the liver image, which plays an extremely important role in the medical field. With the development and innovative application of artificial intelligence algorithms, some studies have used artificial intelligence algorithms for liver image segmentation and feature extraction. [0003] In the existing technology, some studies use the three texture features of entropy local uniformity and gray level...

AI Technical Summary

Problems solved by technology

[0004] However, in the existing technical solutions, there is an obvious shortcoming: when performing feature extraction and segmentation of liver images, it fully relies on the manual and accurate calibration of the original CT image, that is, the manually and accurately calibrated data is required as the segmentation model learning Samples; obviously, a large amount of manual calibration work makes the scale of the learning sample set greatly limited; at the same time, if you want to increase the sample size by a few, you have to spend a lot of human resources of professional doctors, which is time-consuming and labor-intensive, and the feasibility is poor

At the same time, most of the existing technical solutions are based on multi-feature fusion methods, singular value decomposition and wavelet transform methods, which have low feature extraction efficiency, and the information mined is limited, making the extraction and segmentation effects unsatisfactory.

Images