Video Encryption Algorithm With Multiple Levels Based On H.264/SVC Standard

With the rapid growth of video service in web applications recent years, People's requirement have changed from simply chasing video encode standard with high compression ratio to requiring the one that can adapt video stream better to various network environments and user terminals as well as having some fault tolerance and scalability. SVC, Scalable video coding can well solve this problem. SVC can generate video compression code stream with different frame rate and different resolution by only one encoding , and choose quantity of video information that need to be transmitted through network according to different network bandwidth and different capability of display screen as well as terminal decoding, and realize adaptive video quality adjustment. JVT had brought SVC into the extension of H.264/AVC on November 2007.With continuous popularization of scalable video encoding, security problems aimed at scalable video are taken more seriously. Because that scalable video encoding is more complex compare with single layer video encoding, and it also introduced a series of new technology such as temporal scalability, spatial scalability and quality Scalability, so there are more choice in selection of encryption information and design of encryption algorithm.However SVC is so popular now, it has been a short time since SVC appeared, so most of encryption scheme were designed according to technologies in H.264/AVC, for example there were some encryption scheme aimed at motion vector in inter prediction as well as some encryption aimed at DC and AC coefficient in DCT transformation. Because that H.264/AVC standard only refers to single spatial layer encoding, so encryption schemes aimed at H.264/AVC are not suitable for scalable video encoding. On basis of research on H.264/AVC encoding standard and summarizing of past video encryption algorithm, this paper emphatically analyze extensional technology in SVC relative to H.264/AVC standard such as inter-layer prediction in spatial scalability and MGS in quality scalability, and take the lead in putting forward a new video encryption scheme with multiple layer according to characteristic of SVC. This scheme choose distinctive important information of video like MV in inter-layer prediction and reconstruction of key-picture in quality scalability as encryption object, and design algorithm according to the type and characteristic of these information combined with security levels required by users. Face to the situation that application areas of scalable video encoding are continuously enlarging, this scheme can provide different security levels according to different application situation, so that it can guarantee the safety of scalable video when they transmitted through network. Test results show that this scheme has good encryption effect, low volume of secret key and good real-time, and it can also apply to different situations with different security levels.