Page 57 - 2024-Vol20-Issue2
P. 57
53 | Alobaidi & Mikhael
(a) (b) (a) (b)
(c) (d) (c) (d)
(e) (f) (e) (f)
Fig. 4. Sample 3 image from custom-built database/proposed Fig. 5. Sample 4 image from custom-built database/proposed
LSB technique. (a) Original cover image, (b) Stegoimage LSB technique. (a) Original cover image, (b) Stegoimage
with message size of 6KB, (c) Stegoimage with message size with message size of 6KB, (c) Stegoimage with message size
of 8KB, (d) Stegoimage with message size of 10KB, (e) of 8KB, (d) Stegoimage with message size of 10KB, (e)
Stegoimage with message size of 14KB, (f) Stegoimage with Stegoimage with message size of 14KB, (f) Stegoimage with
message size of 16KB. message size of 16KB.
Nevertheless, the smaller window dimensions acquire more VI. CONCLUSIONS
processing resources.
A technique is proposed for inserting a secret message into an
B. Custom-Built Results image, which is based on the two-dimensional Cosine Trans-
As shown in the presented results, the proposed technique form (2D DCT). In this method, the image was converted to
performed better than the other technique in [12]. The RMSE the Cosine domain using 2D DCT, and a predetermined num-
maintained at lower levels while higher PSNRs are achieved. ber of coefficients are chosen to hide the binary secret message.
The visual human inspection illustrate that the proposed tech- The selection process involves analyzing the image in two dif-
nique does not alter the visual properties of the cover image. ferent domains: 2D DCT and 2D Haar Transform. This analy-
In terms of utilized performance metrics, the best block size sis was performed to minimize any distortions in the original
(window dimensions) is 4 × 4 (16 coefficients). Neverthe- cover image. The adaptive algorithm yields weights for each
less, the smaller window dimensions acquire more processing coefficient in its domain, and Cosine coefficients with lower
resources. weights are selected for embedding the secret message. To
evaluate the effectiveness of the technique, samples from the
C. Steganalysis Results BOSSbase, and custom-built databases were used, and three
As shown in Fig. 9, the histogram of the proposed technique metrics were employed: Root Mean Squared Error (RMSE),
has not been altered and thus the proposed technique is im- Sturcutral Similarity Index (SSIM), and Peak Signal-to-Noise
mune against first order attacks like Chi-Square [25]. Ratio (PSNR). Additionally, a visual inspection of the result-
ing image is also taken into account. The results demonstrated
that the proposed technique outperformed commonly used
truncation, energy-based methods, and most recently reported
