Event
Active Sampling for Hardness Classification with Vision-Based Tactile Sensors
1 reports · 1 sources
Junyi Chen and co-authors published a study using an active sampling strategy based on vision-based tactile sensors (VBTS) for sample-efficient hardness classification, selecting samples via information-theoretic measures of model uncertainty. The authors compared three probabilistic classifiers, two active sampling strategies, and a random sampling baseline on a self-built robot-collected VBTS dataset and a published dataset; results show that active sampling methods using variance and entropy as uncertainty measures performed best, achieving 90% accuracy. The study also conducted human hardness classification experiments using the same objects as the dataset, with human accuracy at 48%. The paper has been accepted to the IEEE-RAS International Conference on Humanoid Robots (Humanoids 2026).
Report timeline
- Study: Active sampling with vision-based tactile sensors for hardness classification
The paper investigates active sampling strategies based on vision-based tactile sensors (VBTS) for sample-efficient hardness classification, using information-theoretic measures of model uncertainty. The authors evaluated three probabilistic classifiers, two active sampling strategies, and a random sampling baseline on a hardness classification task using their own VBTS dataset collected by a robot and a previously published dataset, and also conducted a human study on hardness classification with the same objects. Results show that the active sampling methods using variance and entropy as uncertainty measures performed best, reaching 90% accuracy, compared with human accuracy of 48%.
arXiv cs.RO ·