TY  - RPRT
TI  - See and Fix the Flaws: Enabling VLMs and Diffusion Models to Comprehend Visual Artifacts via Agentic Data Synthesis
AU  - Jaehyun Park
AU  - Minyoung Ahn
AU  - Minkyu Kim
AU  - Jonghyun Lee
AU  - Jae-Gil Lee
AU  - Dongmin Park
PY  - 2026
UR  - https://arxiv.org/abs/2602.20951
ID  - 2602.20951
ER  -