BareBones benchmark reveals Vision-Language Models suffer texture bias cliff

Researchers have introduced BareBones, a new benchmark designed to test the geometric comprehension abilities of Vision-Language Models (VLMs). The benchmark uses pixel-level silhouettes to evaluate if VLMs can understand geometric structure independently of visual textures or contextual information. Evaluations of 26 leading VLMs, including GPT-4.1 and Gemini, revealed a significant performance drop when visual textures were removed, a phenomenon termed the "Texture Bias Cliff." AI

影响 Highlights potential limitations in current VLMs' geometric reasoning, suggesting a need for models with better grounding in spatial understanding.

排序理由 The cluster contains a new academic paper introducing a novel benchmark for evaluating Vision-Language Models. [lever_c_demoted from research: ic=1 ai=1.0]

在 arXiv cs.CV 阅读 →

• Aaditya Baranwal
• Claude Sonnet 4.5
• DIS5K
• Gemini
• GPT-4.1
• ImageNet-S
• PASCAL VOC
• ThinObject5K
• WTP-Bench
• LLaVA

• 论文
• 其他

AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →