A Enhancement of Message Capacity Pixel Value Differences (PVD) Steganography Using Filtration

Abstract

Steganography is a widely used technique for securing confidential information by embedding secret messages into digital media, particularly images, in a manner that is imperceptible to human vision. One of the most effective image-based steganography methods is Pixel Value Difference (PVD), which determines embedding capacity based on the pixel intensity differences within each pair of non-overlapping pixels in a cover image. However, in smooth image regions where pixel differences are small, the embedding capacity becomes limited. To address this limitation, several modifications of the PVD technique have been proposed, mainly by increasing the difference values between adjacent pixels before the embedding process. In this research, a simple and practical enhancement approach is introduced by applying image filtering operations to the cover image prior to message insertion. The proposed method utilizes three commonly used filters, namely blur, emboss, and sharpen, with the objective of altering pixel characteristics so that larger pixel value differences can be generated. These filters are selected because they can modify image texture and edge intensity in different ways, which directly influences the data hiding capacity of the PVD algorithm. Experimental results demonstrate that the application of preprocessing filters can significantly improve the performance of PVD steganography in terms of embedding capacity. Furthermore, despite the increased payload, the visual quality of the stego-image remains acceptable, as indicated by favorable Peak Signal-to-Noise Ratio (PSNR) values. This confirms that the sharpen-based enhancement method can improve data hiding efficiency without causing noticeable image distortion. Therefore, the proposed filtering approach offers a simple yet effective solution for increasing the capacity of PVD steganography while preserving image quality.