Sparse Sonomyography-based Estimation of Isometric Force: A Comparison of Methods and Features

• Main Authors: Alzamani, M. (Author), Dockum, A. (Author), Kamatham, A.T (Author), Mukherjee, B. (Author), Sikdar, S. (Author)
• Format: Article
• Language: English
• Published: Institute of Electrical and Electronics Engineers Inc. 2022
• Subjects: Activity sensing; A-mode ultrasound; A-mode ultrasounds; Biomedical signal processing; Clinical research; Electromyography; Estimation; Feature extraction; Features extraction; Force; force estimation; Force estimation; Force predictions; Forecasting; Muscle; Muscle activities; muscle activity sensing.; Muscle activity sensing.; Muscles; Noninvasive medical procedures; Regression analysis; Signal to noise ratio; sonomyography; Sonomyography; Transducers; Ultrasonic imaging; Ultrasonic variable measurement; Ultrasonic variables measurement; Ultrasonography; Wearable technology
• Online Access: View Fulltext in Publisher

 Noninvasive methods for estimation of joint and muscle forces have widespread clinical and research applications. Surface electromyography (sEMG) provides a measure of the neural activation of muscles which can be used to estimate the force produced by the muscle. However, sEMG based measures of force suffer from poor signal-to-noise ratio and limited spatiotemporal specificity. In this paper, we propose a sonomyography or ultrasound imaging-based offline approach for estimating continuous isometric force from a sparse set of ultrasound scanlines. Our approach isolates anatomically relevant features from A-mode scanlines isolated from B-mode images, thus greatly reducing the dimensionality of the feature space and the computational complexity involved in traditional ultrasound-based methods. We evaluate the performance of four regression methodologies for force prediction using the reduced sonomyographic feature set. We also evaluate the feasibility of a practical wearable sonomyography-based system by simulating the effect of transducer placement and varying the number of transducers used in force prediction. Our results demonstrate that Gaussian process regression models outperform other regression methods in predicting continuous force levels from just four equispaced transducers in offline settings. We believe that these findings will aid in the design of wearable, robust and computationally efficient sonomyography-based force prediction systems. IEEE