Research map / Vision-language and multimodal / Human and embodied modeling
ViSioNS: Visual Search in Natural Scenes Benchmark
Fermín Travi, Gonzalo Ruarte, Gaston Bujia, Juan E Kamienkowski
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 IVSN attention map instead of correlation/SSIM as IBS similarity map (nnIBS vs cIBS) (ablation) | IBS · all datasets average | IGhsp: 0.31 -> 0.44 | Improved |
| use IVSN attention map instead of correlation (nnIBS vs cIBS) (ablation) | IBS · all datasets average | LLhsp: 0.23 -> 0.35 | Improved |
| use IVSN attention map instead of SSIM (nnIBS vs sIBS) (ablation) | IBS · all datasets average | IGhsp: 0.31 -> 0.44 | Improved |
| use IVSN attention map instead of correlation/SSIM (ablation) | IBS · all datasets average | AUChsp and NSShsp: 0.75 -> 0.74 | No clear effect |
| use IVSN attention map instead of correlation/SSIM (ablation) | IBS · Multi-Match scanpath similarity | AvgMM: 0.85 -> 0.84 | No clear effect |
| use IVSN attention map instead of correlation (nnIBS vs cIBS) (ablation) | IBS · categorical visual search (COCOSearch18) | ratio of targets found | Improved |
| use SSIM similarity map for template-matching (sIBS) (ablation) | IBS · Interiors template search | ratio of targets found | Improved |
| combine IBS with IVSN attention map (nnIBS) instead of IVSN alone | IVSN · all datasets average | score vs reference: -0.91 -> -0.17 | Improved |
| combine IBS with IVSN attention map (nnIBS) instead of IVSN alone | IVSN · all datasets average | IGhsp: -4.29 -> 0.44 | Improved |
| IVSN instead of nnIBS for search efficiency | nnIBS · all datasets average | AUCperf: 0.55 -> 0.67 | Mixed |
| tune IVSN inhibition-of-return patch size relative to mean target size (ablation) | IVSN · each dataset | search performance | Improved |
| IRL evaluated outside COCOSearch18 | IRL · Interiors and Unrestricted | target found accuracy: 19 of 46; 17 of 42 | Hurt |
Closest papers by meaning
- FlySearch: Exploring how vision-language models explore (NeurIPS 2025 · similarity 0.87)
- CVSearch: Empowering Multimodal LLMs with Cognitive Visual Search for High-Resolution Image Perception (ICML 2026 · similarity 0.86)
- Aligning Vision Models with Human Aesthetics in Retrieval: Benchmarks and Algorithms (NeurIPS 2024 · similarity 0.85)
- Evaluating Multiview Object Consistency in Humans and Image Models (NeurIPS 2024 · similarity 0.85)
- Re-thinking Temporal Search for Long-Form Video Understanding (CVPR 2025 · similarity 0.85)
- EyEar: Learning Audio Synchronized Human Gaze Trajectory Based on Physics-Informed Dynamics (AAAI 2025 · similarity 0.85)
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.
