Mitigating Hallucinations in Multimodal LLMs via Object-aware Preference Optimization

Open in new window