Stochastic BIQA: Median Randomized Smoothing for Certified Blind Image Quality Assessment
This addresses the risk of incorrect quality predictions in public benchmarks for developers of image processing algorithms, though it is incremental as it builds on existing median smoothing with enhancements.
The paper tackled the vulnerability of neural network-based no-reference image quality assessment (NR-IQA) metrics to adversarial attacks by developing a provably robust method using median smoothing and a convolution denoiser with ranking loss, achieving superior SROCC and PLCC scores on three datasets compared to prior methods while maintaining certified guarantees.
Most modern No-Reference Image-Quality Assessment (NR-IQA) metrics are based on neural networks vulnerable to adversarial attacks. Attacks on such metrics lead to incorrect image/video quality predictions, which poses significant risks, especially in public benchmarks. Developers of image processing algorithms may unfairly increase the score of a target IQA metric without improving the actual quality of the adversarial image. Although some empirical defenses for IQA metrics were proposed, they do not provide theoretical guarantees and may be vulnerable to adaptive attacks. This work focuses on developing a provably robust no-reference IQA metric. Our method is based on Median Smoothing (MS) combined with an additional convolution denoiser with ranking loss to improve the SROCC and PLCC scores of the defended IQA metric. Compared with two prior methods on three datasets, our method exhibited superior SROCC and PLCC scores while maintaining comparable certified guarantees.