Dynamic payload adjustment in image steganography through interpolation and center folding strategies

This research introduces a novel interpolation reversible data hiding (IRDH) scheme designed for secure and efficient steganographic techniques, enhancing both image quality and embedding capacity. The interpolation method involves using the original image to zoom in twice for expanding more spaces to embed the secret data. Simultaneously, the values of the original pixels (refer to “seed pixels”) are unchanged that serve for the extraction. Remarkably, the extraction process does not require a reference image, accurately retrieving both the embedded messages and the original image. The embedding procedure divides the cover image into 4 × 4 blocks. Within each block, four seed pixels remain unchanged, serving as references for recovering both the secret data and the cover image. Before embedding the data, the Multi-layer Center Folding Strategy (MCFS) encodes the secret data to minimize distortion. This strategy offers flexibility by allowing dynamic adjustments to the data embedding rate and Peak Signal-to-Noise Ratio (PSNR). The adaptability of this method lies in modifying the bit group parameter ($n$) within MCFS, enabling users to optimize the balance between embedding capacity and image quality. By fine-tuning $n$, the method achieves higher data capacity without compromising the visual integrity of the cover image. Experimental results demonstrate significant improvements in PSNR and embedding capacity compared to existing IRDH techniques.

The highlights of this paper are shown below.

• 1.
  Novel Multi-layer Center Folding Strategy (MCFS): Introduced a new MCFS that significantly reduces image distortion and optimizes the balance between payload capacity and image quality.
• 2.
  Advanced Interpolation-Based Reversible Data Hiding (IRDH) Method: Developed an innovative interpolation technique that enhances embedding capacity and preserves image fidelity, achieving superior PSNR values.
• 3.
  Dynamic Flexibility in Embedding: The proposed method allows adjustable trade-offs between embedding payload and image quality, tailored to specific application requirements.
• 4.
  Improved Security and Robustness: Demonstrated strong resistance to steganalysis attacks (e.g., RS steganalysis and histogram analysis) while maintaining high image quality.
• 5.
  Comprehensive Evaluation: Conducted extensive experiments on multiple datasets, achieving higher PSNR and SSIM values compared to state-of-the-art techniques, showcasing the method's efficacy and generalizability.