VQ-Touch: A Data-Efficient Tactile Generation Framework Across Sensors and Scenarios

2026-07-16

AI Summary

Existing tactile image generation methods rely heavily on large-scale, sensor-specific datasets, resulting in poor generalization and low data efficiency. This work proposes VQ-Touch, a unified generative framework that, for the first time, accommodates mainstream tactile sensors and their variants within a single architecture. The approach leverages a DM-VQGAN to efficiently learn discrete representations of tactile deformations and textures, coupled with a discrete diffusion decoder featuring a unified conditioning interface that enables few-shot, multimodal generation from inputs such as images or labels. Evaluated across diverse sensors and real-world scenarios, VQ-Touch substantially improves both data efficiency and generalization, outperforming state-of-the-art methods on multiple tasks and demonstrating its robustness and effectiveness.

Abstract

Tactile image generation significantly reduces the dependency on expensive and wear-prone sensors by synthesizing high-fidelity tactile data, offering an efficient solution for tactile information acquisition in robotic perception and human-machine interaction systems. However, existing methods depend on large-scale, diverse datasets from specific sensors and lack efficient data utilization and robust generalization capabilities, struggling in vision-limited environments. To address this, we introduce VQ-Touch, a tactile generation framework that supports both cross-sensor and multi-scenario applications. Specifically, to efficiently extract complex deformation and texture features from the data, we propose DM-VQGAN, an effective tactile representation learner. Furthermore, we introduce a discrete diffusion decoder with a unified conditioning interface, supporting multimodal generation tasks such as images and labels, and enhances the model's generalization capability through few-shot mixed training, thus achieving compatibility with current mainstream sensors and their variants. Experiments show that VQ-Touch surpasses state-of-the-art methods in multiple tasks.

Problem

• tactile image generation
• data efficiency
• cross-sensor generalization
• vision-limited environments
• tactile sensing

Innovation

• VQ-Touch
• DM-VQGAN
• discrete diffusion decoder
• cross-sensor tactile generation
• few-shot mixed training