Article Preview
Top1. Introduction
With the rapid development of video coding and network transmission technologies, people can obtain abundant video contents produced by online video service providers such as YouTube, Netflix, Youku, Iqiyi and etc. However, some illegals can easily spread pirated videos through the Internet, which violate the providers’ legal benefits.
In order to protect the security of the video content, video encryption is usually adopted as a safety measure to avoid unauthorized watching. Nevertheless, as the encrypted video is completely unwatchable for customers, the desire of buying the watching service cannot be stimulated. For this reason, many online video providers offer some video highlighting clips or initial part of the video for free watching. It can stimulate consumers paying to watch the complete video to an extent, but it is still unable to give full information about the video to the potential customers. Therefore, researchers have proposed perceptual encryption for this kind of situation (Stutz et al., 2012). The video quality after perceptual encryption is not drastically reduced, and the basic video information can be maintained. After understanding the essential video content, interested customers will pay for the content provider to get the best watching experience.
In recent years, a growing number of ultra-high-definition videos are used in various social fields. Thus, H.264/AVC (Wiegand et al, 2003), which is currently the most widely used video coding standard, cannot meet the requirements of massive videos compression and storage. In 2013, H.265/HEVC was developed as a new generation of video coding standard (Sullivan et al., 2013), which can improve approximately 50% compression ratio compared with H.264. In order to strengthen the security of HEVC videos, a perceptual encryption scheme is proposed in this paper. The main contributions include:
- 1.
A controllable key stream generation method is constructed. By using RC4, a key stream whose proportion of “1” and “0” can be regulated in a delicate designed mechanism, which is used to accomplish the perceptual video encryption.
- 2.
By properly selecting syntax elements for encryption, lossless compression is achieved in the proposed scheme.
- 3.
A perceptual encryption policy is recommended according to the differences of the sensitivity of different syntax elements in encryption.
The rest of the paper is organized as follows. The related work is introduced in Section 2. The perceptual encryption for HEVC videos with lossless compression is presented in Section 3. Experimental results and analysis are provided in Section 4. Finally, some conclusion are drawn in Section 5.
TopThe research of video encryption can be traced back to 1970s, and it became a research hotspot in 1990s. Currently, the main stream of the video encryption is selective encryption, which only encrypts partial data of the video. It can achieve high computational efficiency and format-compliance.
For H.264 encryption, Shahid et al. proposed to encrypt the sign and amplitude of non-zero coefficient in the residual block to scramble video (Shahid et al., 2009, 2011). It can degrade the video quality, but the motion information of the video coding is still kept. Based on the encryption of residual coefficient, intra prediction mode and MVD are also included as the encryption syntax elements by Wang et al. (Wang et al., 2013). It can effectively scramble the texture information and motion information. After the release of HEVC, G. V. Wallendael et al. (Wallendael et al, 2013a,2013b) observed that some syntax elements of HEVC such as MVD, merge index, reference index, residual coefficient and etc., can be selected for encryption, and the encryption of these elements can guarantee format-compliance. Consequently, a format-compliance encryption algorithm for HEVC was proposed in (Wallendael et al, 2014). Recently, Shahid et al. proposed to encrypt the suffix of binarized syntax elements in entropy coding (Shahid et al, 2013). It can keep the size of the encrypted video stream unchanged and has not influence on the compression performance.