Enhanced CNN approaches for multi-image embedding in image steganography

Abstract

In today’s world of information and communication tools, data security is critical for information diffusion. With the growth of extensive multimedia sharing and secret discussions, data concealment has become increasingly vital. Steganography encompasses various types, including image steganography, audio steganography, video steganography, text steganography, network steganography, and digital watermarking. Traditionally, image steganography involves concealing an image within the least significant pixels of a cover image. However, recent advancements have leveraged neural networks to encode and decode secret images within cover images. Our objective is to utilize neural networks especially convolutional neural network to hide multiple images within a single cover image while maximizing payload capacity and minimizing errors in the encoding and decoding processes.