SCAN: Self-Calibrated AutoregressioN for High-Quality Visual Generation
Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (5) · 3976-3984
DOI: 10.1609/aaai.v40i5.37400 ↗Kurz gefasst
Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Zhanzhou Feng und 5 weiteren Autor:innen.
Abstract (Originalsprache)
Human artists can continuously refine their coarse sketches during artistic creation. This is quite different from existing autoregressive generation, where a token is determined once sampled. Aiming to flexibly refine the generated contents, this paper presents a Self-Calibrated AutoregressioN (SCAN) model capable of self-evaluating and refining generation quality without regenerating the entire image. We unify image token generation and quality evaluation into a single autoregressive model, formulating both tasks as categorical prediction problems. During inference, the model first generates a coarse initial image, then iteratively refines the lowest-quality patches until satisfactory image quality is achieved. Experimental results demonstrate that SCAN effectively handles diverse real-world generation errors and achieves a promising balance between image quality and speed. For example, SCAN-XL achieves an FID of 2.10 and an IS of 326.1, surpassing the LlamaGen-XL by 1.29 (+38%) in FID and 99.0 (+43.6%) in IS, with a 5.6× speedup (19.76s to 3.56s). Compared to recent works, SCAN improves FID and speed by +18.3% and +23% over VAR-d20, and by +7% and +46% over RandAR-XL.
Zitieren
APA
Feng, Z., Guo, Q., Chen, J., Gao, F., Yang, M., & Zhang, S. (2026). SCAN: Self-Calibrated AutoregressioN for High-Quality Visual Generation. Proceedings of the AAAI Conference on Artificial Intelligence, 40(5), 3976-3984. https://doi.org/10.1609/aaai.v40i5.37400
Vancouver
Feng Z, Guo Q, Chen J, Gao F, Yang M, Zhang S. SCAN: Self-Calibrated AutoregressioN for High-Quality Visual Generation. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(5):3976-3984. doi:10.1609/aaai.v40i5.37400
BibTeX
@article{Feng2026,
author = {Feng, Zhanzhou and Guo, Qingpei and Chen, Jingdong and Gao, Feng and Yang, Ming and Zhang, Shiliang},
title = {SCAN: Self-Calibrated AutoregressioN for High-Quality Visual Generation},
journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
year = {2026},
volume = {40},
number = {5},
pages = {3976--3984},
doi = {10.1609/aaai.v40i5.37400}
}
EUVIMED