Segmentation method of shoulder musculoskeletal ultrasound structure based on semi-supervised semantic segmentation

Abstract

The invention relates to a shoulder musculoskeletal ultrasonic structure segmentation method based on semi-supervised semantic segmentation, comprising the following steps: preprocessing a plurality of cross-sectional musculoskeletal ultrasonic images obtained to obtain annotated strongly annotated image sets and weakly annotated image sets; using The cross-sectional musculoskeletal ultrasound images in the strongly labeled image set are used to train the generation network and the discriminant network, and then the generation network is reversely optimized using the cross-section musculoskeletal ultrasound images in the weakly labeled image set to obtain the total generation network ; Input the musculoskeletal ultrasound image of the shoulder into the trained total generation network to obtain the ultrasound structure segmentation map of the shoulder musculoskeletal. The invention designs an intelligent recognition method for segmenting musculoskeletal ultrasound images, which saves manpower and time costs in the medical treatment process and improves medical efficiency.

Description

technical field [0001] The invention relates to the technical field of image recognition, in particular to a shoulder musculoskeletal ultrasound structure segmentation method based on semi-supervised semantic segmentation. Background technique [0002] Musculoskeletal ultrasound is an emerging ultrasound examination technology in recent years. The application of high-frequency ultrasound can clearly display superficial soft tissue structures such as muscles, tendons, ligaments, and peripheral nerves. However, due to the large differences in shape and texture between musculoskeletal ultrasound images The problem of the problem causes physicians to spend a lot of time when segmenting and analyzing musculoskeletal ultrasound images, which leads to increased labor costs and time costs. Therefore, being able to design a method for intelligently identifying and segmenting musculoskeletal ultrasound images can save manpower and time costs in the medical process. Contents of the i...

AI Technical Summary

Problems solved by technology

[0002] Musculoskeletal ultrasound is an emerging ultrasound examination technology in recent years. The application of high-frequency ultrasound can clearly display superficial soft tissue structures such as muscles, tendons, ligaments, and peripheral nerves. However, due to the large differences in shape and texture between musculoskeletal ultrasound images The problem caused physicians to spend a lot of time when segmenting and analyzing musculoskeletal ultrasound images, which resulted in increased labor costs and time costs

Images