نسخة أولية وصول مفتوح
SynDORBench: Evaluating LVLM Perceptual Robustness Under Physically Constrained Visibility Conditions
Large vision-language models (LVLMs) have demonstrated remarkable performance on multimodal reasoning benchmarks, yet their perceptual reliability under physically constrained imaging conditions remains poorly understood. Existing evaluations predominantly assume ideal visual inputs and therefore fail to characterize h …