Research map / Detection and segmentation / Open-set and few-shot detection
UniHCP: A Unified Model for Human-Centric Perceptions
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Adapt the pretrained unified model by finetuning on the target dataset | UniHCP · PA-100K | mA: +3.79% | Helped |
| Adapt the pretrained unified model by finetuning on the target dataset | UniHCP · CIHP | mIoU: +4.24% mIoU | Helped |
| Adapt the pretrained unified model by finetuning on the target dataset | UniHCP · CrowdHuman | JI: +1.8% JI | Helped |
| Use task-specific interpreters rather than one shared interpreter (ablation) | UniHCP · In-pretrain datasets (average direct evaluation) | average score: 67.4 -> 67.4 | No clear effect |
| Use task-specific decoders and interpreters instead of sharing them (ablation) | UniHCP · In-pretrain datasets (average direct evaluation) | average score: 67.4 -> 60.6; -6.8% | Hurt |
| Share decoders and interpreters only within each task type (ablation) | UniHCP · In-pretrain datasets (average direct evaluation) | average score: 67.4 -> 65.0; -2.4% | Hurt |
| Use prompt tuning instead of full-model finetuning | UniHCP · ATR (one-shot transfer) | pACC: 90.49 -> 93.65 | Helped |
| Use prompt tuning instead of full-model finetuning | UniHCP · MPII (one-shot transfer) | PCKh: 70.6 -> 83.8 | Helped |
| Use prompt tuning instead of full-model finetuning | UniHCP · PETA (few-shot transfer) | mA: 60.32 -> 62.20 | Helped |
| Add global presence supervision for semantic parts | UniHCP human parsing · Human parsing benchmarks | mIoU: +3.45% mIoU, +3.56% mIoU, and +4.24% mIoU | Helped |
Closest papers by meaning
- A Large-Scale Human-Centric Benchmark for Referring Expression Comprehension in the LMM Era (NeurIPS 2024 · similarity 0.88)
- HumanPCR: Probing MLLM Capabilities in Diverse Human-Centric Scenes (ICLR 2026 · similarity 0.87)
- Human-MME: A Holistic Evaluation Benchmark for Human-Centric Multimodal Large Language Models (ICLR 2026 · similarity 0.86)
- HUNTER: Unsupervised Human-centric 3D Detection via Transferring Knowledge from Synthetic Instances to Real Scenes (CVPR 2024 · similarity 0.86)
- Evaluating Multiview Object Consistency in Humans and Image Models (NeurIPS 2024 · similarity 0.85)
- L-Man: A Large Multi-modal Model Unifying Human-centric Tasks (AAAI 2025 · 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.