Research map / Detection and segmentation / Open-set and few-shot detection
Human-centric Relation Segmentation: Dataset and Solution
Sifeng Liu, Zitian Wang, Yulu Gao, Lejian Ren, Yue Liao, Guanghui Ren, Bo Hu Li, Shuicheng Yan
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 |
|---|---|---|---|
| add multi-level feature fusion on DLA-34 (ablation) | SMS · PIC test, RS | AR: 29.87 -> 30.08 | Improved |
| add multi-level feature fusion on ResNet-50 (ablation) | SMS · PIC test, RS | AR: 29.00 -> 29.97 | Improved |
| add multi-level feature fusion on ResNet-50 (ablation) | SMS · PIC test, HRS | AR: 19.08 -> 19.23 | Improved |
| add multi-level feature fusion on DLA-34 (ablation) | SMS · PIC test, HRS | AR: 19.73 -> 19.73 | No clear effect |
| use SOLOv2 instead of PolarMask for entity segmentation (ablation) | SMS · PIC test, RS | AR: 19.88 -> 29.87 | Improved |
| use SOLOv2 instead of SOLO for entity segmentation (ablation) | SMS · PIC test, HRS | AR: 18.11 -> 19.73 | Improved |
| estimate object offset in addition to subject offset (ablation) | SMS · PIC test, RS | AR: 29.71 -> 29.87 | No clear effect |
| estimate object offset in addition to subject offset (ablation) | SMS · PIC test, HRS | AR: 19.65 -> 19.73 | No clear effect |
| learned interacted part classification instead of most frequent part (ablation) | SMS · PIC test, HRS action | AR: 10.66 -> 12.22 | Improved |
| learned interacted part classification instead of nearest part (ablation) | SMS · PIC test, HRS action | AR: 9.08 -> 12.22 | Improved |
| larger model SMS-L instead of SMS | SMS · PIC test, HRS | AR: 19.23 -> 21.35 | Improved |
| use ResNet-50 instead of DLA-34 backbone | SMS · PIC test, HRS | AR: 19.73 -> 19.23 | Hurt |
Closest papers by meaning
- Hierarchical Human Parsing With Typed Part-Relation Reasoning (CVPR 2020 · similarity 0.88)
- InteractAnything: Zero-shot Human Object Interaction Synthesis via LLM Feedback and Object Affordance Parsing (CVPR 2025 · similarity 0.86)
- Towards Flexible Visual Relationship Segmentation (NeurIPS 2024 · similarity 0.85)
- Human-MME: A Holistic Evaluation Benchmark for Human-Centric Multimodal Large Language Models (ICLR 2026 · similarity 0.85)
- Prototype-Guided Multimodal Relation Extraction based on Entity Attributes (AAAI 2025 · similarity 0.85)
- Human-Object Interaction Detection Collaborated with Large Relation-driven Diffusion Models (NeurIPS 2024 · similarity 0.85)
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.
