TY - RPRT TI - Mask What Matters: Mitigating Object Hallucinations in Multimodal Large Language Models with Object-Aligned Visual Contrastive Decoding AU - Boqi Chen AU - Xudong Liu AU - Jianing Qiu PY - 2026 UR - https://arxiv.org/abs/2602.11737 ID - 2602.11737 ER -