Research map / Vision-language and multimodal / Human and embodied modeling
Self-Training Boosted Multi-Factor Matching Network for Composed Image Retrieval
Haokun Wen, Xuemeng Song, Jianhua Yin, Jianlong Wu, Weili Guan, Liqiang Nie
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 |
|---|---|---|---|
| use multiple latent factor tokens instead of one (U=1) (ablation) | LIMN · FashionIQ-Avg | R@10: 51.77 -> 55.91 | Improved |
| use multiple latent factor tokens instead of one (U=1) (ablation) | LIMN · CIRR | Avg recall: 68.70 -> 72.19 | Improved |
| add batch-based classification loss on latent factor tokens (ablation) | LIMN · FashionIQ-Avg | R@10: 54.43 -> 55.91 | Improved |
| add batch-based classification loss on latent factor tokens (ablation) | LIMN · CIRR | Avg recall: 69.80 -> 72.19 | Improved |
| use CLIP global features to derive latent factor tokens (ablation) | LIMN · FashionIQ-Avg | R@10: 43.92 -> 55.91 | Improved |
| use CLIP global features to derive latent factor tokens (ablation) | LIMN · CIRR | Avg recall: 60.91 -> 72.19 | Improved |
| add filtering of pseudo triplets by LIMN score (ablation) | LIMN+ · FashionIQ-Avg | R@10: 57.17 -> 57.43 | Improved |
| add filtering of pseudo triplets by LIMN score (ablation) | LIMN+ · CIRR | Avg recall: 72.03 -> 72.35 | Improved |
| add iterative dual self-training to LIMN | LIMN · FashionIQ-Avg | R@10: 55.91 -> 57.43 | Improved |
| add iterative dual self-training to LIMN | LIMN · Shoes | R@10: 68.20 -> 68.37 | Improved |
| add iterative dual self-training to LIMN | LIMN · CIRR | R@1: 43.64 -> 43.33 | Hurt |
| add iterative dual self-training to LIMN | LIMN · CIRR | Avg recall: 72.19 -> 72.35 | No clear effect |
Closest papers by meaning
- Language-only Training of Zero-shot Composed Image Retrieval (CVPR 2024 · similarity 0.89)
- Instance-Level Composed Image Retrieval (NeurIPS 2025 · similarity 0.89)
- Visual Delta Generator with Large Multi-modal Models for Semi-supervised Composed Image Retrieval (CVPR 2024 · similarity 0.88)
- Do Composed Image Retrieval Benchmarks Require Multimodal Composition? (NeurIPS 2026 · similarity 0.87)
- ConText-CIR: Learning from Concepts in Text for Composed Image Retrieval (CVPR 2025 · similarity 0.87)
- Learning with Noisy Triplet Correspondence for Composed Image Retrieval (CVPR 2025 · similarity 0.86)
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.
