Resource-Efficient Gesture Recognition through Convexified Attention

TL;DR

This paper presents a convexified attention mechanism for gesture recognition in wearable e-textile devices, achieving high accuracy with significantly fewer parameters and real-time inference, enabling practical on-textile machine learning.

Contribution

Introduces a convexified attention mechanism using Euclidean projection and convex loss for efficient gesture recognition on textile devices, ensuring convergence and minimal resource use.

Findings

Achieved 100% accuracy on tap and swipe gestures

Reduced model parameters by 97% compared to traditional methods

Enabled sub-millisecond inference times

Abstract

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--296$\mu$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.