Tactile MNIST: Benchmarking Active Tactile Perception
Tactile MNIST: Benchmarking Active Tactile Perception
This paper presents the Tactile MNIST Benchmark Suite, an open-source, Gymnasium-compatible framework designed to standardize active tactile perception research. The suite includes diverse simulation scenarios and a comprehensive dataset featuring 13,580 synthetic 3D MNIST digit models and 153,600 real-world tactile samples collected from 600 3D-printed digits. Additionally, the authors train a CycleGAN for realistic tactile simulation rendering, providing standardized protocols and reproducible evaluation frameworks to facilitate systematic progress in tactile sensing and active perception.
The authors introduce an open-source benchmark suite and dataset for active tactile perception, comprising synthetic models and real-world samples collected from 3D-printed digits to support classification, localization, pose estimation, and volume estimation tasks in simulation environments.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
What the source reports
Publisher-reported claims, with original evidence. These results have not been independently verified by RoboSignal.
- Environment
- SimulationOpen source S5
Reported numbers
real-world tactile samples
153,600
View original evidence
ing vision-based tactile sensors. Furthermore, we offer a comprehensive dataset comprising 13,580 synthetic 3D MNIST digit models and 153,600 real-world tactile samples collected from 600 3D-printed digits. Using this dataset, we train a CycleGAN for realistic tactile simulation rendering. By providing standardized protocols and reproducible evaluation frameworks, our benchmark suite facilitates systematic progress i
Open source S5synthetic 3D MNIST digit models
13,580
View original evidence
ing vision-based tactile sensors. Furthermore, we offer a comprehensive dataset comprising 13,580 synthetic 3D MNIST digit models and 153,600 real-world tactile samples collected from 600 3D-printed digits. Using this dataset, we train a CycleGAN for realistic tactile simulation rendering. By providing standardized protocols and reproducible evaluation frameworks, our benchmark suite facilitates systematic progress i
Open source S5printed digits
600
View original evidence
ing vision-based tactile sensors. Furthermore, we offer a comprehensive dataset comprising 13,580 synthetic 3D MNIST digit models and 153,600 real-world tactile samples collected from 600 3D-printed digits. Using this dataset, we train a CycleGAN for realistic tactile simulation rendering. By providing standardized protocols and reproducible evaluation frameworks, our benchmark suite facilitates systematic progress i
Open source S5
- paper: not_reported
- code: not_reported
- weights: not_reported
- dataset: not_reported
What remains unknown
Dataset licensing and commercial-use rights are not established here.
Source excerpts and review record
Automatically extracted; no manual editorial approval recorded.
ing vision-based tactile sensors. Furthermore, we offer a comprehensive dataset comprising 13,580 synthetic 3D MNIST digit models and 153,600 real-world tactile samples collected from 600 3D-printed digits. Using this dataset, we train a CycleGAN for realistic tactile simulation rendering. By providing standardized protocols and reproducible evaluation frameworks, our benchmark suite facilitates systematic progress i
Open source S5
Implications for data suppliers
RoboSignal interpretation and collection questions, not statements of buyer demand.
- Confirm the required data type and collection setting with the buyer; this source does not establish a complete collection specification.
- Compare the reported units and scope before using these quantities in a budget. Recording hours, sensor-hours and trajectories are different measures.
- Validate demand and acceptance criteria with a buyer before scaling. Publication, popularity and a research result do not establish a purchase commitment.
Source：arXiv Robotics — research abstracts · arxiv.org