Table 7 Comparisons with the state-of-the-art on hiding capacity under different priori knowledge and dataset size
From: Coverless image steganography using morphed face recognition based on convolutional neural network
Schemes | Size of dataset | Hiding capacity (bits) | Priori knowledge |
|---|---|---|---|
Zhou et al.’s scheme [27] | 28 = 256 | 8 | None |
Yuan et al.’s scheme [29] | 28 = 256 | 8 | None |
Zhang et al.’s scheme [3] | 215 | 15 (without image division) | Mapping relationship |
Zou et al.’s scheme [32] | 100,000 (hide success rate ≈ 1) | 80 | Chinese dictionary |
Liu et al.’s scheme [34] | 215 | 15 (without image division) | Mapping relationship |
Lu et al.’s scheme [37] | 216 | 16 | Image dataset |
Luo et al.’s scheme [35] | 1 | 2601 | A mapping sequence and a mapping flag; the size of image and image fragments |
Abdulsattar’s scheme [50] | 1 | 6272 | A loop-up table including the location information |
Proposed scheme | 28 | 15 | Image dataset |
300 | 15 | ||
215 | 29 | ||
100,000 | 32 |