Research map / Vision-language and multimodal / Robot action and navigation
TaCo: A Benchmark for Lossless and Lossy Codecs of Heterogeneous Tactile Data
Zhengxue Cheng, Yan Zhao, Keyu Wang, Hengdi ZHANG, Li Song
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| retrain DualComp-I on tactile data and scale to 96M parameters (TaCo-LL-96M) | DualComp-I · TouchandGo lossless | bits/Byte: 0.948 -> 0.447 | Improved |
| retrain DualComp-I on tactile data (TaCo-LL-96M) | DualComp-I · ObjectFolder lossless | bits/Byte: 3.126 -> 2.709 | Improved |
| increase TaCo-LL size from 12M to 96M parameters (ablation) | TaCo-LL · TouchandGo lossless | bits/Byte: 0.622 -> 0.447 | Improved |
| increase TaCo-LL size from 12M to 96M parameters (ablation) | TaCo-LL · ObjTac lossless | bits/Byte: 0.569 -> 0.360 | Improved |
| increase TaCo-LL size from 12M to 96M parameters (ablation) | TaCo-LL · TouchandGo lossless | speed (FPS): 0.67 -> 0.34 | Hurt |
| retrain LALIC on tactile data (TaCo-L) | LALIC · TouchandGo lossy | BD-rate (%): -51.6% -> -61.8% | Improved |
| retrain LALIC on tactile data (TaCo-L) | LALIC · ObjTac lossy | BD-rate (%): 32.8% -> -27.0% | Improved |
| compress with TaCo-L instead of uncompressed data | SVM classifier · TouchandGo material classification | top-1 accuracy (%): 76.63 -> 75.12 | No clear effect |
| compress with TaCo-L instead of uncompressed data | SVM classifier · ObjectFolder material classification | top-1 accuracy (%): 44.11 -> 43.08 | No clear effect |
| compress with TaCo-L instead of LALIC | classification · YCB-Slide SVM | top-1 accuracy (%): 95.67 -> 98.01 | Improved |
| compress with TaCo-L instead of VTM-Intra | grasping · s_lift average | success rate (%): 63.1 -> 62.2 | No clear effect |
| compress with TaCo-L instead of uncompressed tactile signals | grasping · s_lift average | success rate (%): 63.8 -> 62.2 | Hurt |
Closest papers by meaning
- AnyTouch: Learning Unified Static-Dynamic Representation across Multiple Visuo-tactile Sensors (ICLR 2025 · similarity 0.89)
- Tactile MNIST: Benchmarking Active Tactile Perception (NeurIPS 2026 · similarity 0.89)
- Taccel: Scaling Up Vision-based Tactile Robotics via High-performance GPU Simulation (NeurIPS 2025 · similarity 0.88)
- EgoTac: In-the-wild Tactile Prediction from Egocentric Vision (NeurIPS 2026 · similarity 0.88)
- Sensor-Invariant Tactile Representation (ICLR 2025 · similarity 0.87)
- Enhancing Tactile-based Reinforcement Learning for Robotic Control (NeurIPS 2025 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
