Utilizing Imaging Steganographic Improvement using LSB & Image Decoder

Abstract

Hidden data in images, often at the expense of visual quality degradation, are concealed by image steganography. To improve the quality of stego images, this study looks at integrating image decoders and special loss function. Deep learning-based image decoders aid in minimizing distortions caused by embedding, while perceptual loss functions replace traditional metrics to better align with human visual perception. Our goal is to develop a steganographic framework that prioritizes the carrier image's visual integrity while maintaining the hidden data's undetectability. Evaluation will focus on imperceptibility and the objective and subjective quality of the decoded images.