International Journal for Advanced Research in Science & Technology

Image steganography, the art of concealing secret information within an image, has gained significant attention due to its potential applications in secure communication.Traditional steganographic methods often rely on simple transformations of image pixels, which can be easily detected by modern detection algorithms.This project explores a novel approach to image steganography using a Convolutional Neural Network (CNN)-based Encoder-Decoder model.The proposed system leverages deep learning techniques to embed hidden messages in the least significant bits (LSBs) of an image, while simultaneously training an encoderdecoder architecture to efficiently recover the original image and the secret message.The CNN-based encoder-decoder model is designed to optimize both the imperceptibility of the hidden data and the robustness against common image manipulations such as compression or resizing.The encoder extracts features from the input image and encodes the secret message, while the decoder reconstructs the stego-image and decodes the hidden information.To ensure the quality of the stego-image, a loss function is employed that balances the image's perceptual quality and the accuracy of message extraction.