MSCNN-GCN: Graph-Based Tactile Learning Enhancement for Property and Object Recognition With Sparse Sensor Configuration

Tactile perception is essential for autonomous robots to perform intelligent manipulation and effective humanrobot interaction. Unlike visual sensors, tactile sensors can acquire rich physical information through direct contact, enabling recognition of both object properties and categories. However, tactile data are often irregular, heterogeneous, and complex, making effective representation and analysis challenging. In this study, a tactile recognition framework named MSCNN-GCN is proposed. This method combines multiscale convolutional neural networks (CNNs) and graph convolutional networks (GCNs) to enable structured feature extraction from tactile signals. Specifically, a graph is constructed to represent the correlations among tactile channels, and GCNs are used to model their topological and semantic relationships. In parallel, multiscale CNNs extract local spatial features from the tactile input. The two types of features are then integrated using an attention-based fusion mechanism, resulting in enhanced representation capability and improved recognition robustness. Experiments are conducted on a self-collected dataset of nine objects for two tasks, including property recognition and object recognition. MSCNN-GCN achieves accuracies of 93.33% for stiffness recognition, 96.11% for shape recognition, and 95.72% for object recognition, outperforming several baseline methods. The results highlight the effectiveness of GCN-based modeling for tactile data and provide a promising approach for structured learning from multichannel tactile signals.

• Publication year

  2025

• Venue

  IEEE Sensors Journal

• Publication date

  2025-12-15

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1109/JSEN.2025.3623151
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• Spike timing–based coding in neuromimetic tactile system enables dynamic object classification

  2024cited by this paper
• Human tactile sensing and sensorimotor mechanism: from afferent tactile signals to efferent motor control

  2024cited by this paper
• Tactile Object Property Recognition Using Geometrical Graph Edge Features and Multi-Thread Graph Convolutional Network

  2024cited by this paper
• Enhancing Robotic Tactile Exploration With Multireceptive Graph Convolutional Networks

  2024cited by this paper
• Non-destructive fruit firmness evaluation using a soft gripper and vision-based tactile sensing

  2023cited by this paper
• A data-driven robotic tactile material recognition system based on electrode array bionic finger sensors

  2023cited by this paper
• AM-SGCN: Tactile Object Recognition for Adaptive Multichannel Spiking Graph Convolutional Neural Networks

  2023cited by this paper
• MSFR-GCN: A Multi-Scale Feature Reconstruction Graph Convolutional Network for EEG Emotion and Cognition Recognition

  2023cited by this paper
• Machine Learning for Tactile Perception: Advancements, Challenges, and Opportunities

  2023cited by this paper
• Morphology, Connectivity, and Encoding Features of Tactile and Motor Representations of the Fingers in the Human Precentral and Postcentral Gyrus

  2023cited by this paper
• Multi-scale convolutional neural network for texture recognition

  2022cited by this paper
• A non-destructive dropped fruit impact signal imaging-based deep learning approach for smart sorting of kiwifruit

  2022cited by this paper
• Hierarchical Dynamic Graph Convolutional Network With Interpretability for EEG-Based Emotion Recognition

  2022cited by this paper
• Methods to Recognize Depth of Hard Inclusions in Soft Tissue Using Ordinal Classification for Robotic Palpation

  2022cited by this paper
• Multi-Fingered In-Hand Manipulation with Various Object Properties Using Graph Convolutional Networks and Distributed Tactile Sensors

  2022cited by this paper
• Surface Texture Recognition by Deep Learning‐Enhanced Tactile Sensing

  2021cited by this paper
• Visual Perception Enabled Industry Intelligence: State of the Art, Challenges and Prospects

  2021cited by this paper
• Multireceptive Field Graph Convolutional Networks for Machine Fault Diagnosis

  2021cited by this paper
• Interpreting and predicting tactile signals for the SynTouch BioTac

  2021cited by this paper
• Graph Convolutional Network-Based Method for Fault Diagnosis Using a Hybrid of Measurement and Prior Knowledge

  2021cited by this paper
• Fusing Stretchable Sensing Technology with Machine Learning for Human–Machine Interfaces

  2021cited by this paper
• Exploiting dynamic spatio-temporal graph convolutional neural networks for citywide traffic flows prediction

  2021cited by this paper
• A Novel Bilinear Feature and Multi-Layer Fused Convolutional Neural Network for Tactile Shape Recognition

  2020cited by this paper
• Self-Attention Graph Pooling

  2019cited by this paper
• Semi-Supervised Learning With Graph Learning-Convolutional Networks

  2019cited by this paper
• ShuffleNet V2: Practical Guidelines for Efficient CNN Architecture Design

  2018cited by this paper
• T-GCN: A Temporal Graph Convolutional Network for Traffic Prediction

  2018cited by this paper
• GelSight: High-Resolution Robot Tactile Sensors for Estimating Geometry and Force

  2017cited by this paper
• Robotic Tactile Perception of Object Properties: A Review

  2017cited by this paper
• Active Clothing Material Perception Using Tactile Sensing and Deep Learning

  2017cited by this paper
• The role of tactile sensation in online and offline hierarchical control of multi-finger force synergy

  2015cited by this paper
• Deep Residual Learning for Image Recognition

  2015cited by this paper
• Detecting Novel Associations in Large Data Sets

  2011influential reference
• Engineering

  1929cited by this paper

• No citing papers are available for this paper.