AI Paper Atlas

Research map / Vision-language and multimodal / Cross-modal learning

ViSTA: Vision and Scene Text Aggregation for Cross-Modal Retrieval

Mengjun Cheng, Yipeng Sun, Longchao Wang, Xiongwei Zhu, Kun Yao, Jie Chen, Guoli Song, Junyu Han, Jingtuo Liu, Errui Ding, Jingdong Wang

CVPR 2022 · Vision-language and multimodal › Cross-modal learning · approach: Probabilistic embeddings for cross-modal retrieval · 86 citations (OpenAlex)

Check 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 changedWhereMeasuredEffect
Aggregate scene text with image features using ViSTA (ablation)ViSTA-S · CTC-1KImage-to-text R@1: 47.0 -> 52.5Improved
Aggregate scene text with image features using ViSTA (ablation)ViSTA-S · CTC-1KText-to-image R@1: 34.6 -> 36.7Improved
Add image-text contrastive loss alongside fusion-text loss (ablation)ViSTA-S · CTC-1KImage-to-text R@1: 46.6 -> 52.5Improved
Add image-text contrastive loss alongside fusion-text loss (ablation)ViSTA-S · CTC-1KText-to-image R@1: 30.3 -> 36.7Improved
Use fusion-token aggregation rather than global, cross, or late fusion (ablation)ViSTA-S · CTC-1KImage-to-text R@1: 48.4, 50.5, 49.2 -> 52.5Improved
Use fusion-token aggregation rather than global, cross, or late fusion (ablation)ViSTA-S · CTC-1KText-to-image R@1: 34.7, 31.1, 34.9 -> 36.7Improved
Use 4 fusion layers rather than 1 or 2 (ablation)ViSTA-S · CTC-1KImage-to-text R@1: 48.2, 52.2 -> 52.5Improved
Use 4 fusion layers rather than 1 or 2 (ablation)ViSTA-S · CTC-1KText-to-image R@1: 35.6, 35.4 -> 36.7Mixed
Use ViSTA instead of STARNetScene-text-aware retrieval · CTC-1KImage-to-text R@1: 44.1 -> 52.5Improved
Use ViSTA instead of STARNetScene-text-aware retrieval · CTC-1KText-to-image R@1: 31.5 -> 36.7Improved
Use ViSTA-L instead of ViLT-BScene-text-free retrieval · MS-COCO (5K), fine-tunedImage-to-text R@1: 61.5 -> 68.9Improved
Use ViSTA-L instead of ViLT-BScene-text-free retrieval · MS-COCO (5K), fine-tunedText-to-image R@1: 42.7 -> 52.6Improved

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy