TY  - RPRT
TI  - CTCal: Rethinking Text-to-Image Diffusion Models via Cross-Timestep Self-Calibration
AU  - Xiefan Guo
AU  - Xinzhu Ma
AU  - Haiyu Zhang
AU  - Di Huang
PY  - 2026
UR  - https://arxiv.org/abs/2603.20741
ID  - 2603.20741
ER  -