TY - GEN
T1 - On Tiny Feature Engineering: Towards an Embedded EMG-Based Hand Gesture Recognition Model
AU - Bautista, Andres David Gomez
AU - Mendez Chaves, Diego
AU - Alvarado Rojas, Catalina
AU - Mondragón Bernal, Iván Fernando
AU - Colorado, Julian
PY - 2024/12/4
Y1 - 2024/12/4
N2 - Robotic-based therapy is becoming a very popular treatment for the rehabilitation of hand function impairment as a consequence of strokes, due to the high-dimensional and less interpretable nature of superficial electromyography (sEMG) signals. In this context, feature engineering becomes particularly important to estimate the intention of upper limb movements by utilizing machine learning models, specially when a hardware embedded on-board implementation is expected, due to the strong computational, energy and latency constraints. Our work compares the performance achieved by implementing four state-of-the-art feature techniques (random forest, minimum redundancy maximum relevance (MRMR), Davies-Bouldin index, and t-tests), when these are evaluated on a sEMG dataset intended for training hand gesture classifiers. The results of three different machine learning algorithms (neuronal networks, k-nearest neighbors and bagged forest) are used as a reference to validate the analysis. This ongoing research has revealed valuable information on the potential and constraints of these 4 feature generation methods for real gesture recognition embedded applications.
AB - Robotic-based therapy is becoming a very popular treatment for the rehabilitation of hand function impairment as a consequence of strokes, due to the high-dimensional and less interpretable nature of superficial electromyography (sEMG) signals. In this context, feature engineering becomes particularly important to estimate the intention of upper limb movements by utilizing machine learning models, specially when a hardware embedded on-board implementation is expected, due to the strong computational, energy and latency constraints. Our work compares the performance achieved by implementing four state-of-the-art feature techniques (random forest, minimum redundancy maximum relevance (MRMR), Davies-Bouldin index, and t-tests), when these are evaluated on a sEMG dataset intended for training hand gesture classifiers. The results of three different machine learning algorithms (neuronal networks, k-nearest neighbors and bagged forest) are used as a reference to validate the analysis. This ongoing research has revealed valuable information on the potential and constraints of these 4 feature generation methods for real gesture recognition embedded applications.
UR - https://doi.org/10.1109/SEC62691.2024.00049
U2 - 10.1109/SEC62691.2024.00049
DO - 10.1109/SEC62691.2024.00049
M3 - Conference contribution
T3 - 2024 IEEE/ACM Symposium on Edge Computing (SEC)
SP - 437
EP - 442
BT - 2024 IEEE/ACM Symposium on Edge Computing (SEC)
T2 - 2024 IEEE/ACM Symposium on Edge Computing (SEC)
Y2 - 4 December 2024 through 7 December 2024
ER -