Ultrasound imaging of the medial elbow is crucial for the early identification of Ulnar Collateral Ligament (UCL) injuries. Specifically, measuring the elbow joint space in ultrasound images is used to assess the valgus instability of elbow. To automate this measurement, a precisely annotated dataset is necessary; however, no publicly available dataset has been proposed thus far. This study introduces a novel ultrasound medial elbow dataset for measuring joint space to diagnose Ulnar Collateral Ligament (UCL) injuries. The dataset comprises 4,201 medial elbow ultrasound images from 22 subjects, with landmark annotations on the humerus and ulna. The annotations are made precisely by the authors under the supervision of three orthopedic surgeons. We evaluated joint space measurement methods using our proposed dataset with several landmark detection approaches, including ViTPose, HRNet, PCT, YOLOv8, and U-Net. In addition, we propose using Shape Subspace (SS) for landmark refinement in heatmap-based landmark detection. The results show that the mean Euclidean distance error of joint space is 0.116 mm when using HRNet. Furthermore, the SS landmark refinement improves the mean absolute error of landmark positions by 0.010 mm with HRNet and by 0.103 mm with ViTPose on average. These highlight the potential for high-precision, real-time diagnosis of UCL injuries and associated risks, which could be leveraged in large-scale screening. Lastly, we demonstrate point-based segmentation of the humerus and ulna using the detected landmarks as input. The dataset will be made publicly available upon acceptance of this paper at: https://github.com/Akahori000/Ultrasound-Medial-Elbow-Dataset.

翻译：肘关节内侧超声成像对于尺侧副韧带损伤的早期识别至关重要。具体而言，测量超声图像中的肘关节间隙被用于评估肘关节的外翻不稳定性。为了实现该测量的自动化，一个精确标注的数据集是必要的；然而，迄今为止尚未有公开可用的数据集被提出。本研究引入了一个新颖的用于测量关节间隙以诊断尺侧副韧带损伤的超声肘关节内侧数据集。该数据集包含来自22名受试者的4,201张肘关节内侧超声图像，并在肱骨和尺骨上进行了关键点标注。这些标注由作者在三位骨科医生的监督下精确完成。我们使用我们提出的数据集评估了几种关键点检测方法（包括ViTPose、HRNet、PCT、YOLOv8和U-Net）的关节间隙测量性能。此外，我们提出在基于热图的关键点检测中使用形状子空间进行关键点优化。结果显示，使用HRNet时，关节间隙的平均欧几里得距离误差为0.116毫米。此外，形状子空间关键点优化平均将HRNet的关键点位置平均绝对误差改善了0.010毫米，将ViTPose的关键点位置平均绝对误差改善了0.103毫米。这些结果凸显了实现尺侧副韧带损伤及相关风险的高精度、实时诊断的潜力，可应用于大规模筛查。最后，我们展示了以检测到的关键点作为输入，对肱骨和尺骨进行基于点的分割。该数据集将在本文被接受后公开于：https://github.com/Akahori000/Ultrasound-Medial-Elbow-Dataset。