Reversible Data Hiding Algorithm In Encrypted Images And Face Antispoofing Algorithm

In this paper,a separable reversible data hiding method in encrypted images is proposed.The content owner tests an original uncompressed image and generates a location map before encryption.According to the location map,the data-hider embeds additional data into significant bits of the selected pixels with an embedding key.At the receiver side,the original image can be perfectly recovered by a context adaptive interpolation algorithm based on spatial correlation of adjacent pixels.If all additional data are embedded into the most significant bits of the selected pixels,quality of the directly decrypted image is poor;if all additional data are embedded into the second most significant bits,perfect restoration of the original image cannot be achieved.Solving these problems,the proposed method can ensure both good quality of the directly decrypted image and error-less restoration.Experimental results showed feasibility,validity and performance superiority of the method.Regarding another aspect of information security,rendering a face recognition system robust against spoof attacks is of great significance to be dealt with.Aimed at video-based facial spoof attacks,an innovative face antispoofing algorithm based on local binary patterns(LBP)and multiscale discrete cosine transform(DCT)is proposed.First,we extract face images from a target video at a fixed time interval.Second,the low-level descriptors(LBP features)are generated for each extracted face image.After that,we perform multiscale DCT on the low-level descriptors to obtain the high-level descriptors(LBP-MDCT features).In the last stage,the high-level descriptors are fed into a support vector machine(SVM)classifier to determine whether the target video is a spoof attack or not.In contrast to existing approaches,the outstanding experimental results attained by the proposed approach on two widely-used datasets demonstrated its performance superiority,low complexity and high efficiency.