Table 17 Comparison with SOTA using COCO dataset. Significant values are in bold.

From: A deep learning-driven multi-layered steganographic approach for enhanced data security

Method

MSE

PSNR (dB)

SSIM (%)

AccTxt (%)

C (bpp)

Detection rate (%)

Traditional LSB

10.21

52.80

90.75%

78.00%

1.00

89.0%

CNN-Based27

7.15

57.10

93.80%

83.00%

2.00

70.2%

GAN-Based30

6.80

58.00

94.50%

85.50%

2.30

66.0%

Proposed model

5.20

60.50

99.50%

93.80%

2.70

61.5%