Back to feed
AR
arXiv CS.AI
7/21/2026
Design and Validation of a Lightweight 1D CNN for Affective Touch Classification in Soft Plush Companions

Design and Validation of a Lightweight 1D CNN for Affective Touch Classification in Soft Plush Companions

Short summary

This study presents an open-source MATLAB framework for affective touch recognition in soft plush companions, including a FAIR-compliant dataset of 1326 gesture sequences from 25 participants. A compact 13.2k-parameter 1D CNN achieves 75% test accuracy and 85% leave-one-subject-out cross-validation accuracy, with quantized deployment feasible at 20 Hz on target microcontrollers using a hybrid heuristic-plus-CNN pipeline.

  • •Open-source dataset of 1326 labelled gesture sequences from 25 participants for affective touch research
  • •Compact 13.2k-parameter 1D CNN achieves 75% test accuracy, deployable on microcontrollers at 20 Hz
  • •Hybrid heuristic-plus-CNN pipeline proposed for embedded deployment in soft therapeutic companions

Generated with AI, which can make mistakes.

#research-breakthrough#open-source
Read full article at arXiv CS.AI

Is this a good recommendation for you?

Comments

Failed to load comments. Please try again.

Explore more

A Graph Neural Network Model for Real-Time Gesture Recognition Based on sEMG Signals
arXiv CS.AI
Inertia-1: An Open Exploration of Wearable Motion Foundation Models
arXiv cs.LG
Deep Temporal Modeling and Ensemble Fusion for Multimodal Emotion Recognition from Physiological Signals
arXiv cs.CL
Sparse Federated Representation Learning for bio-inspired soft robotics maintenance under real-time policy constraints
Dev.to