@misc{indiciae5fd07abf8f28,
  title = {CLEAR: Unlocking Generative Potential for Degraded Image Understanding in Unified Multimodal Models},
  author = {Xiangzhao Hao and Zefeng Zhang and Zhenyu Zhang and Linhao Yu and Yao Chen and Yiqian Zhang and Haiyun Guo and Shuohuan Wang and Yu Sun},
  year = {2026},
  url = {https://arxiv.org/abs/2604.04780},
  note = {Source identifier: 2604.04780}
}