TY  - RPRT
TI  - CLEAR: Unlocking Generative Potential for Degraded Image Understanding in Unified Multimodal Models
AU  - Xiangzhao Hao
AU  - Zefeng Zhang
AU  - Zhenyu Zhang
AU  - Linhao Yu
AU  - Yao Chen
AU  - Yiqian Zhang
AU  - Haiyun Guo
AU  - Shuohuan Wang
AU  - Yu Sun
PY  - 2026
UR  - https://arxiv.org/abs/2604.04780
ID  - 2604.04780
ER  -