A Novel Video Hiding Method for Biometric Identification
Minghai Yao, Miao Qi · Proceedings of the 2018 5th International Conference on Education, Management, Arts, Economics and Social Science (ICEMAESS 2018) · 2018
In this study, a novel video hiding method is proposed for protecting the transmitted biometric information.Different from the most of existing information hiding methods, the hiding content is an image set rather than a single image or feature vector.For the sake of making the hiding result more invisible and robust, human visual model is adopted by considering both inter-frame and intra-frame motion information.Moreover, the biometric image set is hidden into the video sequence discretely by using discrete wavelet transform method.In particular, in order to detect the integrity of stego-video and guarantee the exactly extraction, the frame number is embedded into the corresponding frame as one watermark.Finally, the extracted image set is used for biometric identification.The identification accuracy is used to verify the performance of hiding method.Experiments show that the proposed hiding method can reach good invisibility and robustness, and ensure the effectiveness of biometrics simultaneously.