A breast ultrasound image classification method based on key point learning

Abstract

The invention discloses a breast ultrasound image classification method based on key point learning. First, the ultrasound image set is preprocessed, the ultrasound image training set is divided into m subsets, and m-1 neural network-based benign and malignant prediction models are trained. ; Then according to the BI-RADS rating standard, the learning method of 6 key points is given, and the multi-level classification prediction of the BI-RADS rating is generated on each model for the input breast ultrasound image; finally, the voting is used to generate the BI-RADS rating Multi-level classification prediction results and adjustment of the corresponding benign and malignant probability values, the present invention solves the technical problem of balancing the effect and interpretability of the current breast ultrasound tumor identification method based on deep learning.

Description

technical field [0001] The invention relates to the field of image processing, in particular to a breast ultrasound image classification method based on key point learning. Background technique [0002] When a breast ultrasound image is given, the doctor will first diagnose the image to determine whether there is a tumor in the image and whether it is a benign or malignant tumor, so as to give a general direction for subsequent treatment. The existing technologies to achieve the same function are roughly as follows: [0003] The first one: based on the medical comprehensive assistance system in commercial software to give identification suggestions. [0004] The general working principle of the medical comprehensive assistance system is as follows: the provider of the commercial medical comprehensive assistance system uses a variety of decision-based processes to determine whether an image is benign or malignant according to its own local data. The software provider's own ...

AI Technical Summary

Problems solved by technology

[0008] In view of the above-mentioned deficiencies in the prior art, a breast ultrasound image classification method based on key point learning provided by the present invention solves the technical problem of balancing the effect and interpretability of the current breast ultrasound tumor recognition method based on deep learning

Images