Visual quality control for perceptual video encryption with multiple transforms

It has been demonstrated in our recent works [1-3] that perceptual video encryption can be effectively achieved by using multiple transforms alternatively according to a random key. In this paper, we introduce a simple but efficient quality control scheme into our encryption algorithm so that the video provider can choose the portion of visual content that is allowed to be disclosed to a un-subscriber for the preview purpose. The proposed scheme can be applied in three different levels during the video encoding: transform coefficients, block, and frame. Experiments are performed using the H.264 standard where two commonly-used quantitative metrics - PNSR and SSIM - are measured. The results show that the degraded visual quality can be controlled linearly with a simple quality factor at each of the three levels. Snap-shots are presented to show that the objective assessment matches the actual video quality.