Research map / Vision-language and multimodal / Cross-modal learning
Cross-Modal Implicit Relation Reasoning and Aligning for Text-to-Image Person Retrieval
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 IRR to the CLIP-ViT-B/16 baseline (ablation) | CLIP-ViT-B/16 baseline · CUHK-PEDES, ICFG-PEDES, and RSTPReid | Rank-1: 68.19 -> 71.23; 56.74 -> 60.96; 54.05 -> 57.90 | Helped |
| Replace baseline InfoNCE with SDM (ablation) | CLIP-ViT-B/16 baseline · CUHK-PEDES, ICFG-PEDES, and RSTPReid | Rank-1: 68.19 -> 70.42; 56.74 -> 60.45; 54.05 -> 57.20 | Helped |
| Replace baseline InfoNCE with CMPM (ablation) | CLIP-ViT-B/16 baseline · CUHK-PEDES and ICFG-PEDES | Rank-1: 68.19 -> 59.31; 56.74 -> 53.83 | Hurt |
| Add ID loss to the SDM+IRR configuration (ablation) | CLIP-ViT-B/16 baseline with SDM and IRR · CUHK-PEDES, ICFG-PEDES, and RSTPReid | Rank-1: 72.81 -> 73.38; 63.27 -> 63.46; 59.25 -> 60.20 | Helped |
| Replace IRRA's interaction module with Co-attn (ablation) | IRRA · CUHK-PEDES | inference time: 6.42 -> 24.30 | Hurt |
| Replace IRRA's interaction module with Merged attn (ablation) | IRRA · CUHK-PEDES | inference time: 6.42 -> 19.20 | Hurt |
| Use IRRA rather than CFine | text-to-image person retrieval · CUHK-PEDES | Rank-1: 69.57 -> 73.38 | Helped |
| Use IRRA rather than CFine | text-to-image person retrieval · ICFG-PEDES | Rank-1: +2.63% | Helped |
| Use IRRA rather than CFine | text-to-image person retrieval · RSTPReid | Rank-1: +9.65% | Helped |
Closest papers by meaning
- Adaptive Uncertainty-Based Learning for Text-Based Person Retrieval (AAAI 2024 · similarity 0.88)
- Chat-based Person Retrieval via Dialogue-Refined Cross-Modal Alignment (CVPR 2025 · similarity 0.88)
- Dynamic Adapter with Semantics Disentangling for Cross-lingual Cross-modal Retrieval (AAAI 2025 · similarity 0.88)
- Beyond Global Similarity: Multi-Conditional Retrieval for Fine-Grained Cross-Modal Understanding (CVPR 2026 · similarity 0.88)
- Adaptive Prompt-Based Semantic Embedding with Inspire Potential of Implicit Knowledge for Cross-Modal Retrieval (AAAI 2025 · similarity 0.87)
- Causality-Inspired Invariant Representation Learning for Text-Based Person Retrieval (AAAI 2024 · 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.