Multi-Image Concealment in High- Resolution Carriers Using Deep Neural Networks

Abstract

The discipline of concealing a private message within an ordinary public message is referred to as steganography. Over time, steganography has utilized techniques such as LSB manipulation and other simple methods to embed images of lower quality into those of higher quality. Our objective is to employ deep neural networks in the process of concealing and revealing multiple concealed images within a single high-resolution cover image. Deep neural networks are designed to operate in tandem, undergoing simultaneous training to enable both the hiding and revealing procedures. Following training on randomly chosen images from the ImageNet database, the system demonstrates effective performance with authentic photographs from a diverse array of origins.