On LSB, LSB OPAP, and LSB OOPAP Published Experimental Results Correctness

Abstract

Recently published (Ansari et al. Iran J Sci Technol Trans Electr Eng 43: 55-65, 2019) experimental results on k least-significant bit (k-LSB) substitution, k-LSB optimal pixel adjustment process (OPAP), and proposed adaptive to an image data-hiding method, k-bit Optimal OPAP (k-LSB OOPAP), do not comply with known to date results. For example, peak signal-to-noise ratio (PSNR) reported is 51.15 dB for 2-LSB OPAP (and for 2-LSB OOPAP) contrary to the known PSNR = 46.27 dB. To resolve the issue, theoretical estimates of mean squared error (MSE) and relying on it PSNR are obtained compliant with more than 99% accuracy with known and our experimental results. Possible reasons for incorrectness of the published experimental results are discussed. Our experiments confirm that k-LSB OPAP and k-LSB OOPAP have practically the same PSNR, and, hence, k-LSB OOPAP is not reasonable as more complicated.