Towards Real-world Image Quality Assessment

Image quality assessment (IQA) is a fundamental task in low-level vision, which has widespread applications in image/video processing, image generation and smart photography, etc. In the literature, a large number of IQA models have been reported with notable achievements. However, the state-of-the-art IQA models are still subject to the generalization challenge when facing real-world problems. In this talk, the latest advances in generalizable image quality assessment will be reviewed, with focus on the distortion diversity and content/theme/scene variation dilemmas. Recent advances on multi-modal IQA will also be discussed.