# Self-supervised perception for tactile skin covered dexterous hands

arXiv: https://arxiv.org/abs/2505.11420 · HTML: https://mlsanity.com/paper/2505.11420/

## Key facts

- Cited by 5 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 16 May 2025.

## Built on by (most cited first, 5 of 5)

- [Dexterity from Smart Lenses: Multi-Fingered Robot Manipulation with In-the-Wild Human Demonstrations](https://mlsanity.com/paper/2511.16661/) (2025)
- [Beyond Binary: Sim-to-Real Dexterous Manipulation with Physics-Grounded Contact Representation](https://mlsanity.com/paper/2605.28812/) (2026)
- [BIDETA: Brain-Inspired Data-Efficient Tactile Adaptation for Unseen Sensors](https://mlsanity.com/paper/2609.08673/) (2026)
- [Tactile-JEPA: Topology-Aware Self-Supervised Representation Learning for Distributed Tactile Sensors](https://mlsanity.com/paper/2609.24385/) (2026)
- [SoTa: Soft Tactile Skins for Dexterous Manipulation](https://mlsanity.com/paper/2610.02338/) (2026)

## Abstract (arXiv)

We present Sparsh-skin, a pre-trained encoder for magnetic skin sensors distributed across the fingertips, phalanges, and palm of a dexterous robot hand. Magnetic tactile skins offer a flexible form factor for hand-wide coverage with fast response times, in contrast to vision-based tactile sensors that are restricted to the fingertips and limited by bandwidth. Full hand tactile perception is crucial for robot dexterity. However, a lack of general-purpose models, challenges with interpreting magnetic flux and calibration have limited the adoption of these sensors. Sparsh-skin, given a history of kinematic and tactile sensing across a hand, outputs a latent tactile embedding that can be used in any downstream task. The encoder is self-supervised via self-distillation on a variety of unlabeled hand-object interactions using an Allegro hand sensorized with Xela uSkin. In experiments across several benchmark tasks, from state estimation to policy learning, we find that pretrained Sparsh-skin representations are both sample efficient in learning downstream tasks and improve task performance by over 41% compared to prior work and over 56% compared to end-to-end learning.