Enhancing Multimodal Large Language Models with Vision Detection Models: An Empirical Study

An empirical study on enhancing MLLMs with state-of-the-art SOTA object detection and Optical Character Recognition models to improve fine-grained understanding and reduce hallucination in responses.