IMAGE STEGANOGRAPHY INSPIRED BY CNN BASED ENCODER-DECODER MODEL

Computer Engineering,

Image steganography is the art of concealing secret information within digital images to ensure covert communication between parties. In recent years, deep learning techniques have shown promising results in various image processing tasks, including steganography. This paper presents an innovative approach to image steganography using a Convolutional Neural Network (CNN) based encoder-decoder model. The proposed model leverages the power of CNNs to learn complex feature representations from images, enabling effective hiding and extraction of secret data. The encoder-decoder architecture is designed to embed the secret information into the cover image seamlessly, while ensuring minimal perceptual distortion. Specifically, the encoder network encodes the secret message into the cover image, producing a stego-image, while the decoder network reconstructs the original message from the stego-image. To enhance the security and robustness of the steganographic system, various techniques such as randomization of embedding positions and adaptive embedding strength are incorporated into the model. Additionally, adversarial training is employed to improve the model's resistance against detection attempts by adversaries. Experimental results demonstrate the effectiveness and robustness of the proposed approach in concealing and recovering secret information while maintaining high visual quality of the stego-images. The proposed CNN-based encoder-decoder model outperforms existing steganographic methods in terms of both embedding capacity and security

This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.