Research map / Video / Video segmentation and tracking
RobotSeg: A Model and Dataset for Segmenting Robots in Image and Video
Haiyang Mei, Qiming Huang, Hai Ci, Mike Zheng Shou
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 |
|---|---|---|---|
| fine-tune SAM 2.1 on robot data (ablation) | SAM 2.1 · VRS-Test whole robot (1C) | J&F: 38.2 -> 73.6 | Improved |
| add cycle consistency loss (ablation) | fine-tuned SAM 2.1 · VRS-Test whole robot (1C) | J&F: 73.6 -> 74.7 | Improved |
| add semantic consistency loss (ablation) | LET · VRS-Test whole robot (1C) | J&F: 74.7 -> 76.3 | Improved |
| add patch consistency loss (ablation) | LET · VRS-Test whole robot (1C) | J&F: 76.3 -> 77.4 | Improved |
| add RPG class tokens (ablation) | RobotSeg · VRS-Test whole robot (AU) | J&F: 77.4 -> 79.8 | Improved |
| add RPG object tokens (ablation) | RobotSeg · VRS-Test whole robot (AU) | J&F: 79.8 -> 83.1 | Improved |
| add SEMA multi-scale perception (ablation) | RobotSeg · VRS-Test whole robot (AU) | J&F: 83.1 -> 83.7 | Improved |
| add SEMA memory-guided structure modulation (ablation) | RobotSeg · VRS-Test whole robot (AU) | J&F: 83.7 -> 85.1 | Improved |
| use RobotSeg instead of RoboEngine finetuned (automatic) | segmentation model · VRS-Test whole robot (AU) | J&F: 80.2 -> 85.1 | Improved |
| use RobotSeg instead of fine-tuned SAM 2.1 on 3-click | segmentation model · VRS-Test robot arm (3C) | J&F: 72.1 -> 76.6 | Improved |
Closest papers by meaning
- Robots Pre-train Robots: Manipulation-Centric Robotic Representation from Large-Scale Robot Datasets (ICLR 2025 · similarity 0.85)
- M^3-VOS: Multi-Phase, Multi-Transition, and Multi-Scenery Video Object Segmentation (CVPR 2025 · similarity 0.85)
- DynaFLIP: Rethinking Robotics Perception via Tri-Modal-Dynamics Guided Representation (NeurIPS 2026 · similarity 0.85)
- A Data-Centric Revisit of Pre-Trained Vision Models for Robot Learning (CVPR 2025 · similarity 0.84)
- ROVER: Recursive Reasoning Over Videos with Vision-Language Models for Embodied Tasks (NeurIPS 2025 · similarity 0.84)
- Robotic Visual Instruction (CVPR 2025 · similarity 0.84)
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.
