Ultrasound Medial Elbow Dataset
Shizuka Akahori and Shotaro Teruya and Pragyan Shrestha and Yuichi Yoshii and Ryuhei Michinobu and Satoshi Iizuka and Akira Ikumi and Hiromitsu Tsuge and Itaru Kitahara

@article{akahori2025ultrasound,
title= {Ultrasound Medial Elbow Dataset},
journal= {},
author= {Shizuka Akahori and Shotaro Teruya and Pragyan Shrestha and Yuichi Yoshii and Ryuhei Michinobu and Satoshi Iizuka and Akira Ikumi and Hiromitsu Tsuge and Itaru Kitahara},
year= {2025},
url= {https://github.com/Akahori000/Ultrasound-Medial-Elbow-Dataset},
abstract= {Ultrasound images of the medial elbow from 22 healthy participants, labeled with eight anatomical landmarks on the humerus and ulna. The dataset comprises 4,201 images according to the paper's abstract. Participants were aged 22 to 56, came from five countries and had no elbow symptoms. Images were captured with a SONIMAGE MX1 (Konica Minolta) and an 11 MHz linear probe, at 528 x 672 pixels and 0.0567 mm per pixel. Two orthopedic surgeons scanned each participant lying down with the elbow flexed to 90 degrees. Landmarks were annotated at sub-pixel precision under the supervision of three orthopedic surgeons. The eight landmarks are the distal end of the humeral trochlea, the proximal end of the ulnar coronoid process, the tip of the humeral medial epicondyle, the deep end of the intersection of the medial epicondyle and the ulnar collateral ligament, the distal end of the medial epicondyle, two points on the hyperechoic lines next to the first two landmarks, and the distal point of the ulnar coronoid process. The first two landmarks define the elbow joint space, which is used to assess valgus instability and ulnar collateral ligament injury. The associated paper benchmarks landmark detectors on it, with HRNet reaching a mean joint space error of about 0.12 mm. Paper: "Measurement of Medial Elbow Joint Space Using Landmark Detection", IEEE Access 2025, DOI 10.1109/ACCESS.2025.3602503. License: CC BY-NC 4.0, non-commercial use only. Please cite the paper.},
keywords= {'medical imaging', 'ultrasound', 'elbow', 'landmark detection'},
terms= {},
license= {https://creativecommons.org/licenses/by-nc/4.0/},
superseded= {}
}