StegaStamp, a novel steganography system designed for practical application in print-scan scenarios. Unlike traditional digital steganography, StegaStamp leverages a generative adversarial network (GAN) architecture to achieve high embedding capacity and imperceptibility while maintaining robustness against common image distortions encountered during printing and scanning. In the initial phase of this project, developed the foundational StegaStamp model using a CNN-based approach. Subsequently, focused on optimizing the model to enhance its performance and efficiency. Through rigorous experimentation and fine-tuning, successfully improved the encoding and decoding speeds, while maintaining a high level of accuracy and preserving the quality of the encoded images. StegaStamp's unique capabilities offer significant advantages over existing steganography techniques. Its ability to seamlessly embed and extract hidden information within printed materials holds immense potential for various applications, including secure authentication, intellectual property protection, and covert communication. By addressing the challenges associated with print-scan scenarios and leveraging the power of GANs, StegaStamp demonstrates a promising solution for practical steganography. Future research directions include exploring further optimizations, expanding its applicability to diverse image formats, and integrating StegaStamp with other security technologies.