Applied Sciences (Aug 2023)
Enhanced Steganography for High Dynamic Range Images with Improved Security and Capacity
Abstract
High-dynamic-range (HDR) images are widely regarded as the ideal format for digital images due to their ability to accurately render a wider range of luminance values. Recently, research has focused on introducing data-hiding techniques to HDR images, but these studies often suffer from a low hiding capacity. In 2011, a steganography scheme was proposed, which utilizes homogeneity in RGBE (red, green, blue, and exponent) format, a popular HDR format, and results in cover images with only slight and ignorable distortions after embedding. However, the capacity of the scheme is limited, and their steganography process may raise suspicions due to the abnormal distribution of pixel values caused by the multiplication and division in the embedding process. There is no denying that security is always the main concern for steganography. A major potential problem became clear after a careful revisiting of the scheme. This paper presents an enhanced steganography scheme that improves embedding capacity by modifying non-embeddable pixels to become embeddable in cover images and avoids potential security weaknesses by using additional random numbers to alter pixel values. The proposed scheme improves the embedding capacity of HDR images while maintaining their visual quality and security against statistical analysis attacks. The experimental result shows that the capacity increases 10 times without visual distortion.
Keywords