An improved image steganography model based on Deep Convolutional Neural Networks

EasyChair Preprint 7705

Abstract

In this paper, we propose an image steganography model with the use of a DeepCNN grounded autoencoder which allows extracting spatiotemporal features from images. It tends to hide four images in one other image taking into consideration equivalency in terms of size. Thus we try to encode and decode multiple secret images within a single, high-resolution cover image. The quantitative result of this model was arranged using the quantitative index "error per pixel", and the qualitative result was evaluated against the existing approaches.

Keyphrases: Auto-encoder, Deep CNN, Steganography, image, spatio-temporal

@booklet{EasyChair:7705,
  author    = {Mounir Telli and Mohamed Othmani and Hela Ltifi},
  title     = {An improved image steganography model based on Deep Convolutional Neural Networks},
  howpublished = {EasyChair Preprint 7705},
  year      = {EasyChair, 2022}}