Research map / Vision-language and multimodal / Cross-modal learning
Integration of Global and Local Representations for Fine-grained Cross-modal Alignment
Seungwan Jin, Hoyoung Choi, Taehyung Noh, Kyungsik Han
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 modality-agnostic adapter (MAA) integrating local patch embeddings (ablation) | FashionSAP-style baseline · FashionGen full evaluation | SumR: 451.82 -> 454.37 | Improved |
| add HNM-F (hard negative mining with focal loss) (ablation) | Baseline + MAA · FashionIQ | SumR: 308.28 -> 318.92 | Improved |
| add C-CmA (comprehensive cross-modal alignment) (ablation) | Baseline + MAA + HNM-F · FashionIQ | SumR: 318.92 -> 323.06 | Improved |
| add C-CmA (ablation) | Baseline + MAA + HNM-F · FashionGen full evaluation | SumR: 456.67 -> 458.11 | Improved |
| add MAA, HNM-F and C-CmA to Fashion-FINE vs FashionSAP | FashionSAP · FashionGen full evaluation ITR R@1 | R@1: 54.43 -> 54.29 | No clear effect |
| add Fashion-FINE modules to BLIP at fine-tuning only (ablation) | BLIP · FashionGen full evaluation | SumR: 451.70 -> 479.87 | Improved |
| add MAA only to BLIP (ablation) | BLIP · FashionGen ITR R@1 | R@1: 55.27 -> 54.60 (ITR R@1); TIR R@1 61.63 -> 65.05 | Mixed |
| weight global vs local representation with lambda 0.95 (ablation) | MAA integration · FashionGen | retrieval performance | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 3 of 3 on most shared results.
- FashionSAP — FashionSAP: Symbols and Attributes Prompt for Fine-Grained Fashion Vision-Language Pre-Training (CVPR 2023) · better on 5 of 6 shared results
- FashionViL — FashionViL: Fashion-Focused Vision-and-Language Representation Learning (ECCV 2022) · better on 3 of 3 shared results
- FashionVLP — FashionVLP: Vision Language Transformer for Fashion Retrieval With Feedback (CVPR 2022) · better on 1 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Easy Regional Contrastive Learning of Expressive Fashion Representations (NeurIPS 2024 · similarity 0.91)
- Cross the Gap: Exposing the Intra-modal Misalignment in CLIP via Modality Inversion (ICLR 2025 · similarity 0.89)
- SyncMask: Synchronized Attentional Masking for Fashion-centric Vision-Language Pretraining (CVPR 2024 · similarity 0.89)
- Linguistic-Aware Patch Slimming Framework for Fine-grained Cross-Modal Alignment (CVPR 2024 · similarity 0.89)
- Same Concept, Different Directions: Cross-Modal Feature Heterogeneity in Sparse Autoencoders (NeurIPS 2026 · similarity 0.88)
- On the Value of Cross-Modal Misalignment in Multimodal Representation Learning (NeurIPS 2025 · similarity 0.88)
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.
