Typing Reinvented: Towards Hands-Free Input via sEMG

TL;DR

This paper presents a novel hands-free text input method using surface electromyography (sEMG) and attention-based models, enabling accurate, real-time typing suitable for VR and spatial computing environments.

Contribution

It introduces an attention-based architecture for sEMG-based typing that outperforms previous convolutional models and demonstrates real-time, accurate text input for wearable interfaces.

Findings

Reduced online CER from 24.98% to 20.34%.

Lowered offline personalized CER from 10.86% to 10.10%.

Achieved fully causal, real-time muscle-driven text input.

Abstract

We explore surface electromyography (sEMG) as a non-invasive input modality for mapping muscle activity to keyboard inputs, targeting immersive typing in next-generation human-computer interaction (HCI). This is especially relevant for spatial computing and virtual reality (VR), where traditional keyboards are impractical. Using attention-based architectures, we significantly outperform the existing convolutional baselines, reducing online generic CER from 24.98% -> 20.34% and offline personalized CER from 10.86% -> 10.10%, while remaining fully causal. We further incorporate a lightweight decoding pipeline with language-model-based correction, demonstrating the feasibility of accurate, real-time muscle-driven text input for future wearable and spatial interfaces.