Towards properties of adversarial image perturbations
This work highlights discrepancies between automated metrics and human perception in image quality assessment, which is important for researchers in computer vision and image processing.
The study investigates adversarial image perturbations that increase the VMAF metric while maintaining subjective quality, finding that moderate brightness changes (~10 pixel units) can boost VMAF by ~60%.
Using stochastic gradient approach we study the properties of adversarial perturbations resulting in noticeable growth of VMAF image quality metric. The structure of the perturbations is investigated depending on the acceptable PSNR values and based on the Fourier power spectrum computations for the perturbations. It is demonstrated that moderate variation of image brightness ($\sim 10$ pixel units in a restricted region of an image can result in VMAF growth by $\sim 60\%$). Unlike some other methods demonstrating similar VMAF growth, the subjective quality of an image remains almost unchanged. It is also shown that the adversarial perturbations may demonstrate approximately linear dependence of perturbation amplitudes on the image brightness. The perturbations are studied based on the direct VMAF optimization in PyTorch. The significant discrepancies between the metric values and subjective judgements are also demonstrated when image restoration from noise is carried out using the same direct VMAF optimization.