A Contrast-Channel Embedding Approach for Image Steganography: Balancing Capacity and Imperceptibility

Steganography, the practice of concealing information within seemingly ordinary media, is essential for protecting sensitive data in today’s digital era. As cyber threats evolve and data transmission grows, ensuring secure and undetectable information hiding becomes increasingly challenging. This study introduces two novel Contrast Channel Embedding (CCE) techniques to optimize steganographic embedding: Mid-Range Contrast Channel Embedding (MCCE) and Extreme-Range Contrast Channel Embedding (ECCE). MCCE targets mid-range contrast areas, avoiding extremes, while ECCE leverages extremely low and high contrast regions, excluding the mid-range. Unlike conventional methods that rely solely on smooth or textured regions, our approach combines gradient-based contrast analysis with RGB contrast thresholds to identify ideal embedding zones. Low-contrast areas enhance imperceptibility by minimizing visible artifacts but limit embedding capacity. In contrast, high-contrast regions offer greater capacity and robustness due to pixel variation, though they must be selected carefully to preserve image quality. We employ Least Significant Bit (LSB) substitution to embed the hidden data efficiently. Experimental results demonstrate that MCCE and ECCE significantly enhance steganographic performance over traditional LSB-based techniques in terms of imperceptibility, maintain a reasonable payload capacity, and improve robustness against detection. Specifically, MCCE achieves PSNR improvements of up to 23.077%, 17.942%, and 12.490%, while ECCE achieves 23.050%, 17.915%, and 12.463%. Both methods also yield consistently higher SSIM scores across various images with a fixed message length of 8192 bytes. These results demonstrate the effectiveness of our approach in providing a comprehensive and secure solution to conceal information in digital images while maintaining visual fidelity.