| With the development of the computer, network and communication technologies, videoconference has gradually become a common and efficient tool for collaborating working and information exchange, and, as a result, has been deployed in many application fields. Since unwanted persons or even opponents may break into the conference system to babble the secrets out, the system is desired to be of high safety. Especially, if the videoconference system is applied in the scenarios of national defense, enterprise and government applications where the national security, military message, business secrets and personal privacy may be involved, the requirements for security are of crucial importance.Due to the intrinsic characteristics of videos such as gigantic data capacity, high redundancy and the additional requirements for video applications like real-time operation and data-format compliance, traditional encryption algorithms are not suitable to videoconference. In this thesis some algorithms intended for video encryption are investigated. Through comparing several existing video encryption schemes and taking the video coding process into account, a general principle was concluded, which requires the algorithms to run fast, be of high security, with low data dilation and of format compliance. Under the above four restrictions, a new partial encryption algorithm was put forward, which encrypts most important data of a video stream like DCT coefficients and motion vectors, to obtain high security with the lowest operation cost. The experimental results have demonstrated the effectiveness of the proposed algorithm.Additionally, a novel partial encryption scheme for facial region is proposed in this thesis. Through a facial tracking algorithm, the presented scheme can automatically track facial regions effectively and the facial regions are protected by employing a simple but secure encryption algorithm. Some encryption examples are shown to testify the availability of the described algorithm. |
PDF Full Text Request