TY  - RPRT
TI  - Better with Less: Tackling Heterogeneous Multi-Modal Image Joint Pretraining via Conditioned and Degraded Masked Autoencoder
AU  - Bowen Peng
AU  - Yongxiang Liu
AU  - Jie Zhou
AU  - Xiaodong Chen
AU  - Tianpeng Liu
AU  - Xiaogang Yu
AU  - Li Liu
PY  - 2026
UR  - https://arxiv.org/abs/2604.16952
ID  - 2604.16952
ER  -