Steganography Algorithm Based on the Nonlocal Maximum Likelihood Noise Estimation for Mobile Applications

Figure 2

Resulting stego-images obtained. Imperceptibility of proposed algorithm via image quality metrics for test images.

(a) PSNR = 52.6092. RMSE = 0.5974. NAE = -0.00006. IF = 0.9999. = 0.9998

(b) PSNR = 51.6823. RMSE = 0.6644. NAE = -0.00004. IF = 0.9999. = 0.9995

(c) PSNR = 38.9523. RMSE = 2.8769. NAE = 0.0002. IF = 0.9985. = 0.9969

(d) PSNR = 39.7741. RMSE = 2.6171. NAE = -0.00009. IF = 0.9986. = 0.9948

(e) PSNR = 40.6538. RMSE = 2.3650. NAE = -0.0001. IF = 0.9986. = 0.9973

(f) PSNR = 46.9981. RMSE = 1.1395. NAE = -0.0025. IF = 0.9851. = 0.9851

(g) PSNR = 51.3264. RMSE = 0.6923. NAE = -0.0001. IF = 0.9998. = 0.9998

(h) PSNR = 41.3803. RMSE = 2.1762. NAE = -0.0001. IF = 0.9988. = 0.9950