End-to-end Trained CNN Encode-Decoder Networks for Image Steganography
This work addresses the need for more efficient and less distorting image steganography methods, which is incremental as it applies deep learning to an existing domain.
The paper tackled the problem of low payload capacity and image distortion in image steganography by proposing an end-to-end trained CNN encoder-decoder architecture, achieving state-of-the-art payload capacity with high PSNR and SSIM values on multiple datasets.
All the existing image steganography methods use manually crafted features to hide binary payloads into cover images. This leads to small payload capacity and image distortion. Here we propose a convolutional neural network based encoder-decoder architecture for embedding of images as payload. To this end, we make following three major contributions: (i) we propose a deep learning based generic encoder-decoder architecture for image steganography; (ii) we introduce a new loss function that ensures joint end-to-end training of encoder-decoder networks; (iii) we perform extensive empirical evaluation of proposed architecture on a range of challenging publicly available datasets (MNIST, CIFAR10, PASCAL-VOC12, ImageNet, LFW) and report state-of-the-art payload capacity at high PSNR and SSIM values.