Estimation of Forearm Pronation–Supination Angles Using MediaPipe and IMU Sensors: Performance Comparison and Interpretability Analysis of Machine Learning Models
Abstract
1. Introduction
2. Materials and Methods
2.1. Participants
2.2. IMU Sensor
2.3. Measurement Protocol
- In-camera: Positioned 30° to the participant’s left front side, 100 cm from the body, and 50 cm above the floor
- Out-camera: Positioned 30° to the participant’s right front side, under the same distance and height conditions
2.4. MediaPipe-Based Data Acquisition and Image Processing
2.5. Parameters
2.6. Machine Learning and Statistical Analysis
- Linear Regression:A fundamental model that assumes a linear relationship between predictors and the target variable. It estimates the coefficients by minimizing the residual sum of squares, offering simplicity and interpretability.
- ElasticNet Regression:A regularized linear regression technique combining L1 (Lasso) and L2 (Ridge) penalties. This allows automatic feature selection while addressing multicollinearity, making it robust in high-dimensional settings.
- Support Vector Machine (SVM):SVM constructs a regression model within a defined error margin and projects data into a higher-dimensional space to capture nonlinearities. It effectively balances complexity and predictive performance.
- Random Forest Regression:An ensemble learning method that aggregates predictions from multiple decision trees trained on bootstrapped datasets with random feature selection. It is robust against overfitting and suitable for nonlinear and hierarchical data patterns.
- LightGBM:Gradient-boosting framework optimized for speed and accuracy. It employs histogram-based learning and exclusive feature bundling, thereby enabling efficient training and high performance on complex large-scale datasets.
- Coefficient of determination (R2): Indicates the proportion of variance in the dependent variable explained by the model. Values closer to 1.0 indicate a better fit.
- MAE: The average of absolute differences between predicted and actual values. It is an intuitive and interpretable error metric.
- Root Mean Square Error (RMSE): The square root of the mean of squared errors. It penalizes larger errors more heavily than MAE.
- Correlation coefficient (r): Measures the strength of the linear relationship between predicted and actual values.
- Feature importance scores, computed from the average gain across all tree splits.
- SHAP (SHapley Additive exPlanations) values [33], a game-theoretic approach for quantifying the impact of each feature on individual predictions.
3. Results
3.1. Validation of Forearm Rotation Angle Using IMU Sensor
3.2. Estimation of Forearm Range of Motion
3.2.1. In-Camera Setting
3.2.2. Out-Camera Setting
3.2.3. Comparison Between Viewpoints
4. Discussion
5. Conclusions
Author Contributions
Funding
Institutional Review Board Statement
Informed Consent Statement
Data Availability Statement
Acknowledgments
Conflicts of Interest
References
- LaStayo, P.C.; Lee, M.J. The forearm complex: Anatomy, biomechanics and clinical considerations. J. Hand Ther. 2006, 19, 137–145. [Google Scholar] [CrossRef]
- Adams, J.E.; Steimmann, S.P. Anatomy and kinesiology of the elbow. In Rehabilitation of the Hand and Upper Extremity, 6th ed.; Skirven, T.M., Osterman, A.L., Fedorczyk, J.M., Amadio, P.C., Eds.; Elsevier: Philadelphia, PA, USA, 2011; Chapter 3; p. 28. [Google Scholar]
- Shen, O.; Chen, C.T.; Jupiter, J.B.; Chen, N.C.; Liu, W.C. Functional outcomes and complications after treatment of distal radius fracture in patients sixty years and over: A systematic review and network meta-analysis. Injury 2023, 54, 110767. [Google Scholar] [CrossRef]
- Soubeyrand, M.; Assabah, B.; Bégin, M.; Laemmel, E.; Dos Santos, A.; Crézé, M. Pronation and supination of the hand: Anatomy and biomechanics. Hand Surg. Rehabil. 2017, 36, 2–11. [Google Scholar] [CrossRef]
- Kapandji, A. Biomechanics of pronation and supination of the forearm. Hand Clin. 2001, 17, 111–122. [Google Scholar] [CrossRef] [PubMed]
- Chen, Y.R.; Tang, J.B. Changes in contact site of the radiocarpal joint and lengths of the carpal ligaments in forearm rotation: An in vivo study. J. Hand Surg. Am. 2013, 38, 712–720. [Google Scholar] [CrossRef]
- Akhbari, B.; Shah, K.N.; Morton, A.M.; Moore, D.C.; Weiss, A.C.; Wolfe, S.W.; Crisco, J.J. Biomechanics of the Distal Radioulnar Joint During In Vivo Forearm Pronosupination. J. Wrist Surg. 2021, 10, 208–215. [Google Scholar] [CrossRef]
- Zhang, H.; Yang, G.; Lu, Y. The role of radial head morphology in proximal radioulnar joint congruency during forearm rotation. J. Exp. Orthop. 2024, 11, e70059. [Google Scholar] [CrossRef]
- Berger, R.; Dobyns, J. Physical examination and provocative maneuvers of the wrist. In Imaging of the Hand; Gilula, L., Ed.; Saunders: Philadelphia, PA, USA, 1996. [Google Scholar]
- Lowe, B.D. Accuracy and validity of observational estimates of wrist and forearm posture. Ergonomics 2004, 47, 527–554. [Google Scholar] [CrossRef]
- Pratt, A.L.; Burr, N.; Stott, D. An investigation into the degree of precision achieved by a team of hand therapists and surgeons using hand goniometry with a standardised protocol. Br. J. Hand Ther. 2004, 9, 116–121. [Google Scholar] [CrossRef]
- Colaris, J.; Van der Linden, M.; Selles, R.; Coene, N.; Allema, J.H.; Verhaar, J. Pronation and supination after forearm fractures in children: Reliability of visual estimation and conventional goniometry measurement. Injury 2010, 41, 643–646. [Google Scholar] [CrossRef] [PubMed]
- Armstrong, A.D.; MacDermid, J.C.; Chinchalkar, S.; Stevens, R.S.; King, G.J. Reliability of range-of-motion measurement in the elbow and forearm. J. Shoulder Elb. Surg. 1998, 7, 573–580. [Google Scholar] [CrossRef] [PubMed]
- Szekeres, M.; MacDermid, J.C.; Birmingham, T.; Grewal, R. The inter-rater reliability of the modified finger goniometer for measuring forearm rotation. J. Hand Ther. 2016, 29, 292–298. [Google Scholar] [CrossRef] [PubMed]
- Yun, X.; Bachmann, E.R.; McGhee, R.B. A simplified quaternion-based algorithm for orientation estimation from earth gravity and magnetic field measurements. IEEE Trans. Instrum. Meas. 2008, 57, 638–650. [Google Scholar] [CrossRef]
- Favre, J.; Aissaoui, R.; Jolles, B.M.; de Guise, J.A.; Aminian, K. A functional calibration procedure for 3D knee joint angle description using inertial sensors. J. Biomech. 2008, 42, 2330–2335. [Google Scholar] [CrossRef]
- Jordan, K.; Dziedzic, K.; Jones, P.W.; Ong, B.N.; Dawes, P.T. The reliability of the three-dimensional fastrak measurement system in measuring cervical spine and shoulder range of motion in healthy subjects. Rheumatology 2000, 39, 382–388. [Google Scholar] [CrossRef]
- Hsu, Y.L.; Wang, J.S.; Lin, Y.C.; Chen, S.M.; Tsai, Y.J.; Chu, C.L.; Chang, C.-W. A wearable inertial-sensing-based body sensor network for shoulder range of motion assessment. In Proceedings of the International Conference on Orange Technologies (ICOT), Tainan, Taiwan, 12–16 March 2013; pp. 328–331. [Google Scholar]
- Parel, I.A.B.; Cutti, A.G.A.; Fiumana, G.C.; Porcellini, G.C.; Verni, G.A.; Accardo, A.P.B. Ambulatory measurement of the scapulohumeral rhythm: Intra- and inter-operator agreement of a protocol based on inertial and magnetic sensors. Gait Posture 2012, 35, 636–640. [Google Scholar] [CrossRef]
- Tan, T.; Chiasson, D.P.; Hu, H.; Shull, P.B. Influence of IMU position and orientation placement errors on ground reaction force estimation. J. Biomech. 2019, 97, 109416. [Google Scholar] [CrossRef]
- Zabat, M.; Ababou, A.; Ababou, N.; Dumas, R. IMU-based sensor-to-segment multiple calibration for upper limb joint angle measurement—A proof of concept. Med. Biol. Eng. Comput. 2019, 57, 2449–2460. [Google Scholar] [CrossRef]
- Pérez-Chirinos Buxadé, C.; Fernández-Valdés, B.; Morral-Yepes, M.; Tuyà Viñas, S.; Padullés Riu, J.M.; Moras Feliu, G. Validity of a Magnet-Based Timing System Using the Magnetometer Built into an IMU. Sensors 2021, 21, 5773. [Google Scholar] [CrossRef]
- Lafayette, T.B.G.; Kunst, V.H.L.; Melo, P.V.S.; Guedes, P.O.; Teixeira, J.; Vasconcelos, C.R.; Teichrieb, V.; da Gama, A.E.F. Validation of Angle Estimation Based on Body Tracking Data from RGB-D and RGB Cameras for Biomechanical Assessment. Sensors 2022, 23, 3. [Google Scholar] [CrossRef] [PubMed]
- Amprimo, G.; Masi, G.; Desiano, M.; Spolaor, F.; De Berti, L.; De Cecco, M. Hand tracking for clinical applications: Validation of the Google MediaPipe Hand (GMH) and the depth-enhanced GMH-D frameworks. Sensors 2024, 24, 584. [Google Scholar] [CrossRef]
- Wagh, V.; Scott, M.W.; Kraeutner, S.N. Quantifying similarities between MediaPipe and a known standard to address issues in tracking 2D upper limb trajectories: Proof of concept study. JMIR Form. Res. 2024, 8, e56682. [Google Scholar] [CrossRef]
- Kusunose, M.; Inui, A.; Nishimoto, H.; Mifune, Y.; Yoshikawa, T.; Shinohara, I.; Furukawa, T.; Kato, T.; Tanaka, S.; Kuroda, R. Measurement of shoulder abduction angle with posture estimation artificial intelligence model. Sensors 2023, 23, 6445. [Google Scholar] [CrossRef] [PubMed]
- Shinohara, I.; Inui, A.; Mifune, Y.; Yamaura, K.; Kuroda, R. Posture Estimation Model Combined With Machine Learning Estimates the Radial Abduction Angle of the Thumb With High Accuracy. Cureus 2024, 16, e71034. [Google Scholar] [CrossRef]
- Ehara, Y.; Inui, A.; Mifune, Y.; Yamaura, K.; Kato, T.; Furukawa, T.; Tanaka, S.; Kusunose, M.; Takigami, S.; Osawa, S.; et al. The development and validation of an artificial intelligence model for estimating thumb range of motion using angle sensors and machine learning: Targeting radial abduction, palmar abduction, and pronation angles. Appl. Sci. 2025, 15, 1296. [Google Scholar] [CrossRef]
- Ogawa, W.; Hirota, Y.; Miyazaki, S.; Nakamura, T.; Ogawa, Y.; Shimomura, I. Definition, criteria, and core concepts of guidelines for the management of obesity disease in Japan (JASSO 2022 Guideline). Endocr. J. 2024, 71, 223–231. [Google Scholar] [CrossRef]
- Pakzad, S.S.; Ghalehnovi, M.; Ganjifar, A. A comprehensive comparison of various machine learning algorithms used for predicting the splitting tensile strength of steel fiber-reinforced concrete. Case Stud. Constr. Mater. 2024, 20, e03092. [Google Scholar] [CrossRef]
- Hassan, M.; Kaabouch, N. Impact of feature selection techniques on the performance of machine learning models for depression detection using EEG data. Appl. Sci. 2024, 14, 10532. [Google Scholar] [CrossRef]
- Enkhbayar, D.; Ko, J.; Oh, S.; Ferdushi, R.; Kim, J.; Key, J.; Urtnasan, E. Explainable artificial intelligence models for predicting depression based on polysomnographic phenotypes. Bioengineering 2025, 12, 186. [Google Scholar] [CrossRef]
- Lundberg, S.M.; Erion, G.; Chen, H.; DeGrave, A.; Prutkin, J.M.; Nair, B.; Katz, R.; Himmelfarb, J.; Bansal, N.; Lee, S.-I. From local explanations to global understanding with explainable AI for trees. Nat. Mach. Intell. 2020, 2, 56–57. [Google Scholar] [CrossRef] [PubMed]
- Matsuki, K.O.; Matsuki, K.; Mu, S.; Sasho, T.; Nakagawa, K.; Ochiai, N.; Takahashi, K.; Banks, S.A. In vivo 3D kinematics of normal forearms: Analysis of dynamic forearm rotation. Clin. Biomech. 2010, 25, 979–983. [Google Scholar] [CrossRef] [PubMed]
- Yoon, H.; Jo, E.; Ryu, S.; Yoo, J.; Kim, J.H. MediaPipe-based LSTM-autoencoder sarcopenia anomaly detection and requirements for improving detection accuracy. In Proceedings of the 1st International Workshop on Intelligent Software Engineering, CEUR Workshop Proceedings, Busan, Republic of Korea, 6 December 2022. [Google Scholar]
Characteristic | Male (n = 10) | Female (n = 10) | Total (n = 20) |
---|---|---|---|
Age, years (mean ± SD, range) | 39.2 ± 6.4 (30–50) | 38.5 ± 6.7 (31–55) | 38.9 ± 6.5 (30–55) |
Body weight, kg (mean ± SD, range) | 69.7 ± 7.8 (60–82) | 59.1 ± 8.1 (48–70) | 64.4 ± 9.5 (48–82) |
BMI, kg/m2 (mean ± SD, range) | 23.7 ± 1.8 (21.0–25.6) | 23.3 ± 3.1 (19.0–26.6) | 23.5 ± 2.5 (19.0–26.6) |
BMI < 25, n (%) | 5 (50%) | 5 (50%) | 10 (50%) |
BMI ≥ 25, n (%) | 5 (50%) | 5 (50%) | 10 (50%) |
Hand dominance, n (%) | Right: 10 (100%) | Right: 10 (100%) | Right: 20 (100%) |
Parameter Name | Definition | Explanation |
---|---|---|
True_angle | Value measured by an IMU sensor | True value |
norm_palm_width | Index–pinky MCP distance/index finger length | |
norm_palm_dist | Wrist–index MCP distance/index finger length | |
forarm_angle | ∠⑤–⓪–E | Angle between index MCP, wrist, and elbow |
rad_palm_angle | ∠⓪–⑤–⑰ | Angle within the hand |
ulnar_palm_angle | ∠⑰–⓪–⑤ | Angle within the hand |
palm_angle | ∠⓪–⑰–⑤ | Angle within the hand |
norm_rad_size | × /2 | Index–pinky–wrist triangle area/(index finger length)2 |
norm_ulnar_size | × /2 | Pinky–wrist–index triangle area/(index finger length)2 |
norm_palm_size | × /2 | Wrist–index- pinky triangle area/(index finger length)2 |
Direction | MAE (°) | SD (°) |
---|---|---|
Pronation 90° | 1.29 | 0.96 |
Pronation 60° | 2.26 | 2.55 |
Pronation 30° | 2.23 | 2.59 |
Neutral (0°) | 1.88 | 2.07 |
Supination 30° | 3.13 | 2.97 |
Supination 60° | 3.26 | 1.03 |
Supination 90° | 7.13 | 1.46 |
Linear Regression | scikit-learn 1.3 |
|
Elastic-Net | scikit-learn 1.3 |
|
SVM | scikit-learn 1.3 |
|
Random Forest Regression | scikit-learn 1.3 |
|
Light GBM | LightGBM 4.2 |
|
MAE (°) | RMSE (°) | R2 | Correlation Coefficient | |
---|---|---|---|---|
Linear Regression | 14.25 | 17.99 | 0.851 | 0.922 |
ElasticNet | 14.25 | 17.99 | 0.851 | 0.922 |
SVM | 6.25 | 8.57 | 0.966 | 0.983 |
Random Forest Regression | 6.83 | 8.83 | 0.964 | 0.982 |
LightGBM | 5.61 | 7.64 | 0.973 | 0.986 |
MAE (°) | RMSE (°) | R2 | Correlation Coefficient | |
---|---|---|---|---|
Linear Regression | 21.03 | 27.62 | 0.660 | 0.813 |
ElasticNet | 21.18 | 28.24 | 0.645 | 0.803 |
SVM | 5.44 | 8.03 | 0.971 | 0.986 |
Random Forest Regression | 5.83 | 7.84 | 0.973 | 0.986 |
LightGBM | 4.65 | 7.30 | 0.976 | 0.988 |
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content. |
© 2025 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).
Share and Cite
Kusunose, M.; Inui, A.; Mifune, Y.; Yamaura, K.; Shinohara, I.; Tanaka, S.; Ehara, Y.; Takigami, S.; Osawa, S.; Nakabayashi, D.; et al. Estimation of Forearm Pronation–Supination Angles Using MediaPipe and IMU Sensors: Performance Comparison and Interpretability Analysis of Machine Learning Models. Appl. Sci. 2025, 15, 10527. https://doi.org/10.3390/app151910527
Kusunose M, Inui A, Mifune Y, Yamaura K, Shinohara I, Tanaka S, Ehara Y, Takigami S, Osawa S, Nakabayashi D, et al. Estimation of Forearm Pronation–Supination Angles Using MediaPipe and IMU Sensors: Performance Comparison and Interpretability Analysis of Machine Learning Models. Applied Sciences. 2025; 15(19):10527. https://doi.org/10.3390/app151910527
Chicago/Turabian StyleKusunose, Masaya, Atsuyuki Inui, Yutaka Mifune, Kohei Yamaura, Issei Shinohara, Shuya Tanaka, Yutaka Ehara, Shunsaku Takigami, Shin Osawa, Daiji Nakabayashi, and et al. 2025. "Estimation of Forearm Pronation–Supination Angles Using MediaPipe and IMU Sensors: Performance Comparison and Interpretability Analysis of Machine Learning Models" Applied Sciences 15, no. 19: 10527. https://doi.org/10.3390/app151910527
APA StyleKusunose, M., Inui, A., Mifune, Y., Yamaura, K., Shinohara, I., Tanaka, S., Ehara, Y., Takigami, S., Osawa, S., Nakabayashi, D., Higashi, T., Wakamatsu, R., Hayashi, S., Matsumoto, T., & Kuroda, R. (2025). Estimation of Forearm Pronation–Supination Angles Using MediaPipe and IMU Sensors: Performance Comparison and Interpretability Analysis of Machine Learning Models. Applied Sciences, 15(19), 10527. https://doi.org/10.3390/app151910527