Secured Data Hiding and Sharing using Improved LSB-based Image Steganography Technique

Despite the notable performance of various existing image steganography techniques in the spatial domain, various statistical analyses such as Chi-square and RS analysis still considerably affect their over-all performance. Also, employing conventional cryptosystems to further secure the secret message adds significant overhead to the computational complexity. This paper presents a less complex and straight-forward image steganography technique that has an acceptable embedding capacity, an appropriate level of security and a significantly low probability of detection of either visual assessment or statistical analysis techniques. In this paper, the embedding process is not directly executed to the original color space of the cover image which is the RGB but rather in YCbCr color space. The Huffman encoding helped to attain the high payload capacity and the concept of quadratic residues strengthened the additional layer of security. The PSNR and SSIM results showed that after a 100 percent embedding to the LSBs of the cover image, distortions were within the satisfactory range. Also, the Chi-square and RS analysis simulation results proved that the stego-images generated by the proposed technique have a low probability of detection.