Parameter optimized edge guided image steganography using variable bit LSB embedding for high capacity and visual fidelity

Secure communication has attracted many researchers attentioin in recent years, but steganography techniques still face the challenges to balance out the payload capacity, imperceptibility, and robustness. The proposed work introduced here is Canny edge detection based on parameter variation of low and high thresholds and variance. Unlike conventional approaches that adopt fixed settings, the proposed method explores the parameters to identify saturation points where edge density stabilizes, ensuring reproducibility and efficiency. A variable-bit LSB substitution mechanism is then applied, embedding more bits in edge pixels and fewer in non-edge regions, thereby exploiting the masking properties of the human visual system., and reproducible embedding strategy, making it well suited for applications in secure multimedia communication, digital forensics, and data protection in applied engineering domains. Extensive experiments on benchmark grayscale images demonstrate that the proposed method achieves payloads up to 3.3 bpp, with peak-signal-to-noise (PSNR) values exceeding 34 dB and structural similarity index measure (SSIM) above 0.82, thus ensuring competitive imperceptibility. Comparative study with other prior developed techniques shows significant improvement in terms of performance evaluation parameters.