A Robust DWT–PVD Image Steganography for Secure Data Embedding

Abstract

The purpose of image steganography, a common information-hiding method, is to embed hidden data into digital images while retaining their visual quality and preventing statistical detection. Although Discrete Wavelet Transform (DWT)-based methods improve imperceptibility by using the frequency-domain characteristics, they frequently have a limited payload capacity. Conventional Pixel Value Differencing (PVD) techniques offer high embedding capacity but may introduce noticeable distortion in smooth image regions. This study suggests a novel mshDWTPVD model that combines the benefits of the Pixel Value Differencing (PVD) method and the Haar Discrete Wavelet Transform (DWT) in order to get around these restrictions. The most suitable sub-band is selected for data embedding, followed by coefficient quantization using different quantization factors (Q = 2, 4, 6, and 8). The quantized coefficients are shifted to positive values, and secret message bits are embedded adaptively using the PVD technique based on pixel difference ranges. Finally, inverse quantization and inverse DWT are performed to reconstruct the stego image. Both the quantization factor and the embedding sub-band can be changed in order to evaluate the proposed model. The statistical parameters are among the objective quality metrics used to analyze performance. The proposed mshDWTPVD model achieves good visual quality, great embedding capacity and computing efficiency, according to experimental data. Additionally, the comparison analysis shows that choosing a suitable high-frequency sub-band and quantization factor substantially enhances imperceptibility while preserving strong embedding performance, making the proposed approach feasible for reliable and secure image steganography applications.