Encryption Edge Points in Face using Light Stream Algorithm

Unauthorized sharing and data theft represent a huge challenge for everyone, especially for companies and sensitive governmental institutions, where the videos might be hacked or viewed by untheorized parties. The video encryption can protect the video content from the outsiders. However, a full video encryption consumes time; thus, a partial encryption is used for efficient and quick encryption. In this paper, a partial encryption has been presented. It is feasible to use one of the video cryptography algorithms, where only the edges in the face area are encrypted in the video frames. Also, a light encryption was used in the present research for the video encryption, and the Salsa20 algorithm was utilized for the face edge encryption, while Viola-Jones was employed for the face area detection. The Canny Edge detection operator retrieves the regions of interest from the face detection. PSNR (31.856), MSE (42.402), correlation, UACI, NPCR, and entropy measures were used for evaluating the proposal. The percentages of comparison ratio manifested that the proposed algorithm consumed less time (1.033 s) for the encryption and outperformed other algorithms in many evaluation metrics, such as decryption time and entropy of the decrypted image.