Saved in:
Bibliographic Details
Main Authors: Han, Isaac, Lee, Seoyoung, Park, Sangyeon, Akan, Ecehan, Luo, Yiyue, DelPreto, Joseph, Kim, Kyung-Joong
Format: Preprint
Published: 2026
Subjects:
Online Access:https://arxiv.org/abs/2603.25906
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912984278761472
author Han, Isaac
Lee, Seoyoung
Park, Sangyeon
Akan, Ecehan
Luo, Yiyue
DelPreto, Joseph
Kim, Kyung-Joong
author_facet Han, Isaac
Lee, Seoyoung
Park, Sangyeon
Akan, Ecehan
Luo, Yiyue
DelPreto, Joseph
Kim, Kyung-Joong
contents Estimating human pose, classifying actions, and predicting movement progress are essential for human-robot interaction. While vision-based methods suffer from occlusion and privacy concerns in realistic environments, tactile sensing avoids these issues. However, prior tactile-based approaches handle each task separately, leading to suboptimal performance. In this study, we propose a Shared COnvolutional Transformer for Tactile Inference (SCOTTI) that learns a shared representation to simultaneously address three separate prediction tasks: 3D human pose estimation, action class categorization, and action completion progress estimation. To the best of our knowledge, this is the first work to explore action progress prediction using foot tactile signals from custom wireless insole sensors. This unified approach leverages the mutual benefits of multi-task learning, enabling the model to achieve improved performance across all three tasks compared to learning them independently. Experimental results demonstrate that SCOTTI outperforms existing approaches across all three tasks. Additionally, we introduce a novel dataset collected from 15 participants performing various activities and exercises, with 7 hours of total duration, across eight different activities.
format Preprint
id arxiv_https___arxiv_org_abs_2603_25906
institution arXiv
publishDate 2026
record_format arxiv
spellingShingle Shared Representation for 3D Pose Estimation, Action Classification, and Progress Prediction from Tactile Signals
Han, Isaac
Lee, Seoyoung
Park, Sangyeon
Akan, Ecehan
Luo, Yiyue
DelPreto, Joseph
Kim, Kyung-Joong
Computer Vision and Pattern Recognition
Estimating human pose, classifying actions, and predicting movement progress are essential for human-robot interaction. While vision-based methods suffer from occlusion and privacy concerns in realistic environments, tactile sensing avoids these issues. However, prior tactile-based approaches handle each task separately, leading to suboptimal performance. In this study, we propose a Shared COnvolutional Transformer for Tactile Inference (SCOTTI) that learns a shared representation to simultaneously address three separate prediction tasks: 3D human pose estimation, action class categorization, and action completion progress estimation. To the best of our knowledge, this is the first work to explore action progress prediction using foot tactile signals from custom wireless insole sensors. This unified approach leverages the mutual benefits of multi-task learning, enabling the model to achieve improved performance across all three tasks compared to learning them independently. Experimental results demonstrate that SCOTTI outperforms existing approaches across all three tasks. Additionally, we introduce a novel dataset collected from 15 participants performing various activities and exercises, with 7 hours of total duration, across eight different activities.
title Shared Representation for 3D Pose Estimation, Action Classification, and Progress Prediction from Tactile Signals
topic Computer Vision and Pattern Recognition
url https://arxiv.org/abs/2603.25906