AI Paper Atlas

Research map / Vision-language and multimodal / Multimodal LLMs

Vision-Language Models Can't See the Obvious

Ngoc Dung Huynh, Phuc H Le-Khac, Wamiq Reyaz Para, Ankit Singh, Sanath Narayan

ICCV 2025 · Vision-language and multimodal › Multimodal LLMs · approach: VLM capability analysis · 20 citations (OpenAlex)

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 changedWhereMeasuredEffect
use synthetic instead of natural imagesGPT-4o · SalBench Detection 5-shotF1: 41.9 -> 86.0Improved
scale model size from 1.5B to 72BQwen2-VL · SalBench Detection synthetic 5-shotF1: 23.8 -> 54.9 -> 89.9Improved
scale model size from 1.5B to 72BQwen2-VL · SalBench Detection natural 5-shotF1: 25.3 -> 37.2 -> 43.9Improved
add few-shot examples (0 to 5 shots)GPT-4o · SalBench Detection naturalF1: 47.6 -> 38.9 -> 41.9Hurt
add few-shot examples (0 to 5 shots)Phi3.5-Vision-3.5B · SalBench Visual Referring syntheticF1: 63.5 -> 19.0Hurt
add few-shot examples (0 to 5 shots)Qwen2-VL-72B · SalBench Visual Referring syntheticF1: 74.7 -> 87.9Improved
increase number of distractors (<7 to >25)Qwen2-VL-72B · SalBench natural Detectionavg F1: 44.5 -> 37.4Hurt
increase number of distractors (<7 to >25)GPT-4o · SalBench natural Detectionavg F1: 47.0 -> 41.6Hurt
increase size difficulty (easy to hard)Qwen2-VL-72B · SalBench synthetic Size Detectionaccuracy: 94.2 -> 46.0Hurt
increase size difficulty (easy to hard)GPT-4o · SalBench synthetic Size Detectionaccuracy: 93.3 -> 36.8Hurt
increase color difficulty (medium to hard)Qwen2-VL-72B · SalBench synthetic Color Detectionaccuracy: 100.0 -> 60.1Hurt
increase orientation difficulty (easy to hard)Qwen2-VL-72B · SalBench synthetic Orientation Visual Referringaccuracy: 88.3 -> 71.7Hurt

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