Strengthening wavelet based image steganography using Rubik’s cube segmentation and secret image scrambling
Jyoti Khandelwal,Vijay Kumar Sharma
DOI: https://doi.org/10.1007/s11042-024-18576-3
IF: 2.577
2024-02-29
Multimedia Tools and Applications
Abstract:Image steganography is an extremely rich and significant exploration region that gives productive answers to some genuine and modern issues. This paper deals with secret image transmission and securing it from different attacks. The proposed image steganography technique uses a combination of image segmentation, pre-processing, and scrambling methos. The proposed technique begins with the segmentation of secret and cover images using Rubik’s cube. The segmentation process breaks the images into six parts, then after they are stored in a segmented image dataset. In the second step, each segmented secret image is processed by the image pre-processing method. The proposed image pre-processing method helps the system select the appropriate cover image from the segmented cover image dataset, which makes the technique more robust. In the third step, before the embedding process, secret images are scrambled using the proposed scrambling method. Finally, the cover and secret images are processed by discrete wavelet transforms (DWT) and singular value decomposition (SVD) to generate the stego image. The proposed technique is tested on a variety of grayscale images, with peak signal-to-noise ratio (PSNR) values ranging from 32.27 to 30.77 (dB) and structural similarity index measure (SSIM) values ranging from 0.93 to 0.90 for different alpha values. The extracted secret image is analyzed using normalized correlation (NC) and naturalness image quality evaluator (NIQE) parameters. The NC values are above 0.99, and the NIQE values ranged from 3.3264 to 3.8468 for various extracted images. To analyze the proposed technique, measured value of entropy and an elapse time are 7.7956 and 6.489 s, respectively. Comparative studies are conducted in four main areas. The first comparative study tested the reliability of the proposed scrambling method by calculating the overall and diagonal correlation values of the secret image and scrambled secret image. The second comparative study tested the impeccability and reliability of the stego image by comparing its PSNR value to other stego images produced by other researchers. The third comparative study tested the effect of false-positive tests on extracted secret images. The results showed that the extracted secret image could only be obtained from the original stego image, indicating that the proposed method was secure against false positives. The fourth or last comparative study compares the computational load of proposed technique with previously developed techniques. In general, the results of the comparative study show that the proposed technique is reliable and secure for image scrambling, stego image generation, and secret image extraction.
computer science, information systems, theory & methods,engineering, electrical & electronic, software engineering