Liver tumor automatic segmentation positioning method and system and storage medium

Abstract

The invention discloses a liver tumor automatic segmentation positioning method and system and a storage medium, The method comprises: performing window width window processing on the liver CT image to obtain the first input data of the highlighted CT liver part; inputting the first input data into the U-Net liver semantic segmentation model to obtain the liver semantic segmentation result; filtering the segmentation result to obtain the second input data; reconstructing the second input data to obtain the third input data; inputting the third input data into the tumor location detection modelto obtain the tumor location information; and acquiring the tumor location information according to the tumor location information; filtering the second input data to obtain the fourth input data; inputting the fourth input data into the U-Net tumor semantic segmentation model to obtain the tumor semantic segmentation result. The invention can quickly obtain the result of tumor semantic segmentation, reduce the time cost, improve the work efficiency, and improve the accuracy of the tumor detection, and can be widely applied to the field of image processing technology.

Description

technical field [0001] The invention relates to the technical field of image processing, in particular to a method, system and storage medium for automatic segmentation and positioning of liver tumors. Background technique [0002] Explanation of terms: [0003] Deep learning: The concept of deep learning originated from the research of artificial neural networks. Deep learning is a branch of machine learning and a method based on representational learning of data. Deep learning combines low-level features to form more abstract high-level representation attribute categories or features to discover the distributed feature representation of data. Therefore, unlike traditional machine learning, which allows researchers to manually construct data features, deep learning does not require manual construction. feature. Like machine learning methods, deep machine learning methods are also divided into supervised learning and unsupervised learning, and the learning models establish...

AI Technical Summary

Problems solved by technology

If the area of ​​liver tissue resected is too large, it will affect the postoperative recovery of the patient, but if the lesion remains after resection, the patient will have the risk of recurrence. Therefore, liver tumor detection and contour drawing are the key tasks for the smooth progress of treatment one

The traditional manual image reading method relies on a lot of time cost and the experience of radiologists, and manual image reading is prone to errors and the accuracy is not high. Therefore, an automatic detection and segmentation method for liver tumors is very important for clinical practice.

Images