Liver tumor segmentation method and device based on multi-stage CT image guidance

Abstract

The embodiment of the invention provides a liver tumor segmentation method and device based on multi-stage CT image guidance, and the method comprises the steps: obtaining an abdomen CT image with enhanced contrast, inputting the abdomen CT image with enhanced contrast into a preset single-channel full convolutional neural network, and obtaining a liver region-of-interest image; inputting the liver region-of-interest image into a preset tumor segmentation network to obtain a tumor segmentation result corresponding to the contrast-enhanced abdominal CT image; wherein the tumor segmentation network is obtained by performing multi-channel fusion training according to liver region-of-interest image samples in different periods and image samples with tumor region marks corresponding to the abdominal CT image samples. According to the embodiment of the invention, the liver tumor CT images in different periods are effectively subjected to feature mining by adopting a multi-channel fusion training network, so that the trained network has higher segmentation precision and better robustness on liver tumors.

Description

technical field [0001] Embodiments of the present invention relate to the technical field of medical image processing, and more specifically, to a method and device for segmenting liver tumors based on multi-phase CT image guidance. Background technique [0002] Automatic and accurate segmentation of liver tumors from abdominal CT sequence images is very important for many liver-related clinical operations. In recent years, with the continuous maturity of CT imaging technology, it is helping doctors in the clinical examination and diagnosis of liver cancer has been widely applied. However, segmenting liver tumors from abdominal CT images has always is a difficult problem. In the traditional diagnosis and treatment of liver tumors, doctors need to rely on their rich experience and professional knowledge to judge the location of the liver and its tumors on each slice. In the case of a large number of patients, it will lead to difficulties in queuing numbers and a high rate ...

AI Technical Summary

Problems solved by technology

[0004] There are large differences in tumor size, shape, location, and number between different patients, and the tumor boundary is not clear, which brings great difficulties to traditional segmentation methods

Segmentation methods based on machine learning often require manual design of feature extraction methods for tumors. The design of feature extraction methods directly affects the final segmentation performance, and the technical barriers are high.

Training liver tumor segmentation models with deep learning methods requires a large amount of data. Most of the existing methods are trained based on image data collected during unenhanced scans, and such data can reflect very limited information about liver tumors. Indistinct, even a professional doctor can hardly judge

Images