An Optimized Selective Encryption for Video Confidentiality

Selective Encryption (SE) offers effective and efficient protection of visual content for assuring video confidentiality. However, it is challenging to optimize SE’s computational overhead while maintaining encryption effectiveness. This paper suggests the study of classifying the targeted sign-flip Quantized Coefficients (QC) so as to minimize the Encryption Computation Overhead (ECO), while maximizing Perceptual Encryption Effectiveness (PEE) both across the H.264/AVC and the HEVC video coding standards. The evaluation of this min-max optimization is based on the number of encryption bits used to flip the targeted QC sign and the Peak Signal to Noise Ratio (PSNR) estimated post encryption for quantifying ECO and PEE respectively. Results of simulation on different video sequences indicate that the derived set of QC’s can reduce ECO cost by up to 78 % without affecting the video confidentiality when compared to state-of-the-art approaches.