Research map / Vision-language and multimodal / Human and embodied modeling
Web-Based Semantic Fragment Discovery for On-Line Lingual-Visual Similarity
Xiaoshuai Sun, Jiewei Cao, Chao Li, Lei Zhu, Heng Tao Shen
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 CNN fc7 features instead of GIST (ablation) | self-similarity quality assessment · Flickr30K-Quality | AUC: 0.7376 -> 0.8537 | Improved |
| use fc6 relu instead of fc6 features (ablation) | self-similarity measure · Flickr30K-Quality | AUC: 0.8512 -> 0.8512 | No clear effect |
| use fc7 relu instead of fc7 (ablation) | self-similarity measure · Flickr30K-Quality | AUC: 0.8537 -> 0.8525 | No clear effect |
| increase object proposals from 1 to 20 (option P) (ablation) | label-free retrieval · Flickr30K | Recall@10: 33.5 -> 37.1 | Improved |
| increase object proposals from 1 to 20 (option T) (ablation) | label-free retrieval · Flickr30K | Recall@10: 26.7 -> 28.3 | Improved |
| increase object proposals from 1 to 20 (option G) (ablation) | label-free retrieval · Flickr30K | Recall@10: 33.7 -> 37.4 | Improved |
| use option P instead of T (20 OP) (ablation) | label-free retrieval · Flickr30K | Recall@1: 7.4 -> 10.4 | Improved |
| use option G instead of P (20 OP) (ablation) | label-free retrieval · Flickr30K | Recall@10: 37.1 -> 37.4 | No clear effect |
| use Eqn. 4 instead of Eqn. 5 option G (20 OP) (ablation) | label-free retrieval · Flickr30K | Recall@1: 10.5 -> 10.7 | No clear effect |
| use fc7 relu instead of fc6 CNN layer (ablation) | label-free retrieval · Flickr30K | Recall@10: 28.4 -> 37.4 | Improved |
| use fc7 instead of fc6 relu CNN layer (ablation) | label-free retrieval · Flickr30K | Recall@10: 33.6 -> 33.9 | No clear effect |
| increase top object proposals from 1 to 10 (ablation) | label-free retrieval · Flickr30K | Recall@10: 31.0 -> 36.7 | Improved |
Closest papers by meaning
- MERIT: Multilingual Semantic Retrieval with Interleaved Multi-Condition Query (NeurIPS 2025 · similarity 0.86)
- EvdCLIP: Improving Vision-Language Retrieval with Entity Visual Descriptions from Large Language Models (AAAI 2025 · similarity 0.85)
- Asymmetric Visual Semantic Embedding Framework for Efficient Vision-Language Alignment (AAAI 2025 · similarity 0.85)
- Meta CLIP 2: A Worldwide Scaling Recipe (NeurIPS 2025 · similarity 0.85)
- Visual-Text Cross Alignment: Refining the Similarity Score in Vision-Language Models (ICML 2024 · similarity 0.85)
- The Many Senses of Visual Similarity: A Text-Prompted Image Perceptual Metric (NeurIPS 2026 · similarity 0.84)
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.
