Keypoint detection and functional evaluation of human lower limbs based on YOLOv8 and HRNet
6 affiliations
- 1The 2nd Ward of Joint Surgery Department, Tianjin University Tianjin Hospital
- 2College of Exercise Science & Health, Tianjin University of Sport
- 3Rehabilitation Department, Tianjin Hospital
- 4College of Artificial Intelligence, Hebei University of Technology
- 5The 2nd Ward of Joint Surgery Department, Tianjin Hospital
- 6School of Information Engineering, Harbin Institute of Petroleum
Abstract
With the development of artificial intelligence and deep learning technology, human pose estimation has been widely applied in fields such as medical rehabilitation and motion analysis. To meet the needs of objective assessment of lower limb function in patients with knee joint diseases, we propose a method for detecting and assessing the function of human lower limb keypoints by integrating YOLOv8 with an improved HRNet. Firstly, the YOLOv8s model is used to locate the human region in the image. Then, the HRNet-W32 model, enhanced with a Gating Unit and Keypoint Attention Unit, is employed to detect six lower limb keypoints: left/right hip, left/right knee, and left/right ankle. Finally, based on the geometric relationship of hip-knee-ankle, the flexion-extension angle of the knee joint is calculated, and the sitting-to-standing action is recognized through a threshold state machine. Experimental results show that the improved HRNet achieves a 3.9% and 1.3% increase in [email protected] and [email protected] metrics, respectively, compared to the original HRNet-W32. In 200 sets of 30-Second Sit-to-Stand Test videos, the system achieves an average counting accuracy of 97.84% and an average absolute error of 0.18 times. The correlation between the visual algorithm output and clinical functional status was further evaluated using established clinical indicators, including the Knee injury and Osteoarthritis Outcome Score (KOOS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Timed Up and Go (TUG) test, and the 30-Second Sit-to-Stand Test.
Related
- Optimization of ⁶⁸Ga-lactosylated human serum albumin (LSA) for PET-based quantitative evaluation of hepatic functional reserve in liver fibrosis models.
- Evaluation of the BL-RED (Beta-Lactamase Rapid Electrochemical Detection) test for the rapid detection of resistance to third-generation cephalosporins in Enterobacteriaceae
- Evaluation of Structural and Functional Outcomes After the Application of an Initial Single Dose of Faricimab
- Perfusion and neuro-functional evaluation of new methods for protection from ischemia–reperfusion injury in microvascular reconstructive post-oncologic surgery
- Evaluation of iodine detection and quantification limits in photon-counting CT—a phantom study
BibTeX
@article{yang2026keypoint,
title = {{Keypoint detection and functional evaluation of human lower limbs based on YOLOv8 and HRNet}},
author = {Tao Yang and Yichi Zhang and Jing Wang and Wen Luo and Dezhi Zheng and Yanmin Bu and Jianan Li and Ruoying Li and Lei Wang and Huiwen Zhao},
journal = {Scientific Reports},
year = {2026},
month = jul,
doi = {10.1038/s41598-026-62913-5},
url = {https://doi.org/10.1038/s41598-026-62913-5},
note = {CC BY},
}
RIS
TY - JOUR
TI - Keypoint detection and functional evaluation of human lower limbs based on YOLOv8 and HRNet
AU - Tao Yang
AU - Yichi Zhang
AU - Jing Wang
AU - Wen Luo
AU - Dezhi Zheng
AU - Yanmin Bu
AU - Jianan Li
AU - Ruoying Li
AU - Lei Wang
AU - Huiwen Zhao
PY - 2026
DA - 2026/07/01
JO - Scientific Reports
DO - 10.1038/s41598-026-62913-5
UR - https://doi.org/10.1038/s41598-026-62913-5
AB - With the development of artificial intelligence and deep learning technology, human pose estimation has been widely applied in fields such as medical rehabilitation and motion analysis. To meet the needs of objective assessment of lower limb function in patients with knee joint diseases, we propose a method for detecting and assessing the function of human lower limb keypoints by integrating YOLOv8 with an improved HRNet. Firstly, the YOLOv8s model is used to locate the human region in the image. Then, the HRNet-W32 model, enhanced with a Gating Unit and Keypoint Attention Unit, is employed to detect six lower limb keypoints: left/right hip, left/right knee, and left/right ankle. Finally, based on the geometric relationship of hip-knee-ankle, the flexion-extension angle of the knee joint is calculated, and the sitting-to-standing action is recognized through a threshold state machine. Experimental results show that the improved HRNet achieves a 3.9% and 1.3% increase in [email protected] and [email protected] metrics, respectively, compared to the original HRNet-W32. In 200 sets of 30-Second Sit-to-Stand Test videos, the system achieves an average counting accuracy of 97.84% and an average absolute error of 0.18 times. The correlation between the visual algorithm output and clinical functional status was further evaluated using established clinical indicators, including the Knee injury and Osteoarthritis Outcome Score (KOOS), Western Ontario and McMaster Universities Osteoarthritis Index (WOMAC), Timed Up and Go (TUG) test, and the 30-Second Sit-to-Stand Test.
SN - 2045-2322
ER -