Research map / Detection and segmentation / CNN features
Devil in the Details: Towards Accurate Single and Multiple Human Parsing
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 |
|---|---|---|---|
| Add high-resolution embedding (ablation) | CE2P baseline on LIP · LIP validation | mIoU: 47.97 -> 49.11 | Helped |
| Add global context embedding (ablation) | CE2P baseline on LIP · LIP validation | mIoU: 47.97 -> 49.44 | Helped |
| Add edge perceiving (ablation) | CE2P baseline on LIP · LIP validation | mIoU: 47.97 -> 50.04 | Helped |
| Add edge perceiving to global-context and high-resolution modules (ablation) | CE2P on LIP · LIP validation | mIoU: 51.54 -> 52.56 | Helped |
| Use left-right flipping at inference (ablation) | CE2P on LIP · LIP validation | mIoU: 52.56 -> 53.10 | Helped |
| Use three-model ensemble with different input sizes instead of a single model (ablation) | CE2P on LIP · LIP test | mIoU: 56.50 -> 57.90 | Helped |
| Fuse local parsing branches with the global-image branch (ablation) | M-CE2P on MHP v2.0 · MHP v2.0 validation | AP^p_0.5: 28.71 -> 30.92 | Helped |
| Apply label refinement (ablation) | M-CE2P on MHP v2.0 · MHP v2.0 validation | AP^p_0.5: 30.92 -> 34.47 | Helped |
| Add global-image parsing to both local branches (ablation) | M-CE2P on MHP v2.0 · MHP v2.0 validation | AP^p_0.5: 28.71 -> 30.92 | Helped |
| Use global-image parsing branch alone rather than local parsing (ablation) | M-CE2P on MHP v2.0 · MHP v2.0 validation | AP^p_0.5: 24.04 | Hurt |
| Use CE2P instead of JPPNet | LIP validation · LIP validation | mIoU: 51.37 -> 53.10 | Helped |
Closest papers by meaning
- Look Into Person: Self-Supervised Structure-Sensitive Learning and a New Benchmark for Human Parsing (CVPR 2017 · similarity 0.91)
- Correlating Edge, Pose With Parsing (CVPR 2020 · similarity 0.89)
- A Large-Scale Human-Centric Benchmark for Referring Expression Comprehension in the LMM Era (NeurIPS 2024 · similarity 0.84)
- A Pedestrian is Worth One Prompt: Towards Language Guidance Person Re-Identification (CVPR 2024 · similarity 0.83)
- Human-MME: A Holistic Evaluation Benchmark for Human-Centric Multimodal Large Language Models (ICLR 2026 · similarity 0.83)
- An Empirical Study of CLIP for Text-Based Person Search (AAAI 2024 · similarity 0.83)
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.