PERFORMANCE EVALUATION OF STEGANOGRAPHY TECHNIQUES ACROSS COLOUR MODELS USING HUMAN VISION METRICS

Abstract

Steganography is the field of study that focuses on the various approaches for hiding data in such a manner that, its existence cannot be noticed. This is achieved by hiding information in innocent looking objects suitable for covert communication. Different methods have been proposed for hiding secret data within various types of cover media eg texts, images, videos, audios and even computer network resources. Image Steganography is the most popular method for secretly conveying messages especially because it comes with the advantage of possessing capacity of very large space for information hiding purposes. Additionally, several colour models exist for use in steganography. Colour images are stored in 24 bits depth while grey scale images are stored in 8 bits. Since existing colour models have varying capacities and efficiencies for information hiding, determining which colour model to use during covert communication remains a major challenge in the research community. Therefore, this work carried out the implementation of the Least Significant Bit (LSB) and Discrete Cosine Transformation (DCT) steganography technique, and also carried out its performance evaluation across colour models using human vision standard metrics. Results obtained from our experiments reveal that, Hue-Saturation-Value (HSV) colour model is the most efficient colour model for information hiding, when compared with other colour spaces such as the Red-Green-Blue (RGB), Luminance-Chroma Blue-Chroma Red (YCbCr) and the Lightness-Green-red-Blue-yellow (L*a*b) colour spaces. This is because, HSV exhibited high inserting capacity and also better stego – image as evident in the fact that, HSV returned greater PSNR (Peak Signal to Noise Ratio values) which implies higher quality of stego-image, after embedding secret information. Overall, our study established the strength, robustness, suitability and security level of these colour spaces for information hiding.