Logo image
Back
Resource-Efficient Gesture Recognition through Convexified Attention
Journal article   Open access   Peer reviewed

Resource-Efficient Gesture Recognition through Convexified Attention

Daniel Schwartz, Dario Salvucci, Yusuf Osmanlioglu, Richard Vallett, Genevieve Dion and Ali Shokoufandeh
Proceedings of the ACM on human-computer interaction, v 10(4), pp 1-25
29 Jun 2026
DOI: https://doi.org/10.1145/3815369
url
https://doi.org/10.1145/3815369View
Published, Version of Record (VoR) Open

Abstract

Applied computing Classification and regression trees Computing methodologies Consumer products Convex optimization Empirical studies in HCI Evaluation General and reference Gestural input Hardware Health informatics Human-centered computing Neural networks Regularization Sensor devices and platforms Sensors and actuators Supervised learning Tactile and hand-based interfaces Theory of computation Touch screens Ubiquitous and mobile computing theory, concepts and paradigms Ubiquitous and mobile devices
Wearable e-textile interfaces require gesture recognition capabilities but face severe constraints in power consumption, computational capacity, and form factor that make traditional deep learning impractical. While lightweight architectures like MobileNet improve efficiency, they still demand thousands of parameters, limiting deployment on textile-integrated platforms. We introduce a convexified attention mechanism for wearable applications that dynamically weights features while preserving convexity through nonexpansive simplex projection and convex loss functions. Unlike conventional attention mechanisms using non-convex softmax operations, our approach employs Euclidean projection onto the probability simplex combined with multi-class hinge loss, ensuring global convergence guarantees. Implemented on a textile-based capacitive sensor with four connection points, our approach achieves 100.00% accuracy on tap gestures and 100.00% on swipe gestures—consistent across 10-fold cross-validation and held-out test evaluation—while requiring only 120–360 parameters, a 97% reduction compared to conventional approaches. With sub-millisecond inference times (290–296μ s) and minimal storage requirements (< 7KB), our method enables gesture interfaces directly within e-textiles without external processing. Our evaluation, conducted in controlled laboratory conditions with a single-user dataset, demonstrates feasibility for basic gesture interactions. Real-world deployment would require validation across multiple users, environmental conditions, and more complex gesture vocabularies. These results demonstrate how convex optimization can enable efficient on-device machine learning for textile interfaces.

Metrics

1 Record Views

Details

Logo image