EUVIMEDEUVIMED entdecken
Crossref

Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment

Ziheng Jia, Jiaying Qian, Zicheng Zhang, Zijian Chen, Xiongkuo Min

Proceedings of the AAAI Conference on Artificial Intelligence · 14.03.2026 · 40 (27) · 22301-22309

DOI: 10.1609/aaai.v40i27.39387 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Proceedings of the AAAI Conference on Artificial Intelligence, online erschienen am 14.03.2026. Verfasst von Ziheng Jia und 4 weiteren Autor:innen. Die Arbeit wurde bisher 8-mal zitiert (Crossref).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Reinforcement fine-tuning (RFT) is a proliferating paradigm for LMM training. Analogous to high-level reasoning tasks, RFT is similarly applicable to low-level vision domains, including image quality assessment (IQA). Existing RFT-based IQA methods typically use rule-based output rewards to verify the model's rollouts but provide no reward supervision for the "think” process, leaving its correctness and efficacy uncontrolled. Furthermore, these methods typically fine-tune directly on downstream IQA tasks without explicitly enhancing the model’s native low-level visual quality perception, which may constrain its performance upper bound. In response to these gaps, we propose the multi‐stage RFT IQA framework (Refine-IQA). In Stage-1, we build the Refine-Perception-20K dataset (with 12 main distortions, 20,907 locally-distorted images, and over 55K RFT samples) and design multi-task reward functions to strengthen the model’s visual quality perception. In Stage-2, targeting the quality scoring task, we introduce a probability difference reward involved strategy for "think" process supervision. The resulting Refine-IQA Series Models achieve outstanding performance on both perception and scoring tasks—and, notably, our paradigm activates a robust "think” (quality interpretating) capability that also attains exceptional results on the corresponding quality interpreting benchmark.

Zitieren

APA

Jia, Z., Qian, J., Zhang, Z., Chen, Z., & Min, X. (2026). Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment. Proceedings of the AAAI Conference on Artificial Intelligence, 40(27), 22301-22309. https://doi.org/10.1609/aaai.v40i27.39387

Vancouver

Jia Z, Qian J, Zhang Z, Chen Z, Min X. Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment. Proceedings of the AAAI Conference on Artificial Intelligence. 2026;40(27):22301-22309. doi:10.1609/aaai.v40i27.39387

BibTeX
@article{Jia2026,
  author = {Jia, Ziheng and Qian, Jiaying and Zhang, Zicheng and Chen, Zijian and Min, Xiongkuo},
  title = {Refine-IQA: Multi-Stage Reinforcement Finetuning for Perceptual Image Quality Assessment},
  journal = {Proceedings of the AAAI Conference on Artificial Intelligence},
  year = {2026},
  volume = {40},
  number = {27},
  pages = {22301--22309},
  doi = {10.1609/aaai.v40i27.39387}
}
Kostenlos starten →