In-depth learning of automatic segmentation of shoulder joint magnetic resonance images based on convolutional neural networks

Artificial intelligence-based applications in shoulder surgery leaves much to be desired: a systematic review

“…Mu et al used approximately 800 MRI images to test and evaluate the performance of AI models in automatically segmenting different bony regions of interest from shoulder MRIs. 46 Similarly, Wang et al also used approximately 800 MRI images to develop and test an AI model for automated detection, classification, and segmentation of bone regions of interest in shoulder MRIs. 64 Likewise, Rodrigues et al used over 100 patients to apply AI models for fully automated segmentation of the glenohumeral joint and quantification of glenoid anatomy, glenoid bone loss, and humeral anatomy using shoulder MRIs.…”

Artificial intelligence powered advancements in upper extremity joint MRI: A review

“…An average verification accuracy of 94% was achieved on the humerus and shoulder girdle segmentation on a minimal dataset. Building on this, the algorithm has been integrated into the medical image measurement and analysis platform "3DQI", through which the 3D segmentation effect of shoulder joint bones can be displayed, and it can provide clinical diagnosis guidance to orthopedics [ 54 ].…”

A review of advances in image-guided orthopedic surgery

“…They introduced an enhanced dual attention module and a two-class feature fusion module to improve the accuracy of the segmentation. For the segmentation of the shoulder joint in MRI, Mu et al (2021) proposed a model which combines U-Net and Alex Net. This model firstly used three U-Nets and an Alex Net to focus on an area, and then used Alex Net to refine the segmentation.…”