Search

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.

Fe2O3 nanoparticles of different sizes ranging from 22 to 56nm were synthesized chemically by a modified sol-gel method. Pure alpha phase particles as well as particles with admixture of alpha and gamma phase were obtained and identified by x-ray and M\"ossbauer measurements. Different size and phase controlling parameters have been identified. The average size of the particles decreases with increased annealing temperature of the gel and decreases with the increase in the concentrations of the (citric acid). The annealing temperature affects the relative fractions of the two phases and consequently the magnetization of the particles. The magnetization of the particles (Ms) and coercivity (Hc) increase consistently with the increase in the annealing temperature. For the same relative amount of the two phases, the coercivity (Hc) and the moment of the particles increase with the decrease in the size of the particles, indicating the role of surface effects in the magnetic behaviour., Comment: 19 pages, PDF file