Novel multi-media steganography model using meta-heuristic and deep learning assisted adaptive lifting wavelet transform

Abstract

The major goal of this proposal is to implement a new steganography model with multi-media data. Initially, in video steganography and frames are extracted and then, the keyframes with texture and motion features of gathered video are selected through Deep Convolutional Neural Networks (DCNN). Then, the optimal region selection for embedding is carried out using the Improved Artificial Gorilla Troops Optimizer (IAGTO). Secondly, image steganography is performed by collecting the images, where the color channel selection is done through DCNN. Thirdly, audio steganography is performed by gathering the input audios from the standard datasets. Further, the optimal time instant selection is performed by the same IAGTO algorithm, which will be fed to the next phase. Finally, the audio embedding is done by ALWT, and then, the stego audios are extracted by taking inverse ALWT. From the experimental analysis, successful embedding over multi-media data is achieved..