VCU-Bridge: Hierarchical Visual Connotation Understanding via Semantic Bridging

Ming Zhong, Yuanlei Wang, Liuzhou Zhang +7

While Multimodal Large Language Models (MLLMs) excel on benchmarks, their processing paradigm differs from the human ability to integrate visual information. Unlike humans who natu…