2021/01/03 by Rohit Agrawal, Agrawal, Rohit, Kapil Ahuja +1
Computer Science · #Advanced Steganography and Watermarking Techniques #Chaos-based Image/Signal Encryption #Digital Media Forensic Detection #E.3 #FOS: Computer and information sciences #FOS: Mathematics #I.4.2 #I.4.5 #I.4.9 #Multimedia (cs.MM) #Optimization and Control (math.OC)
paper · pdf · doi:10.48550/arxiv.2101.00690
openalex publication_date 2021/01/03 · openalex created_date 2022/07/25 · openalex updated_date 2026/07/28
Image steganography plays a vital role in securing secret data by embedding\nit in the cover images. Usually, these images are communicated in a compressed\nformat. Existing techniques achieve this but have low embedding capacity.\nEnhancing this capacity causes a deterioration in the visual quality of the\nstego-image. Hence, our goal here is to enhance the embedding capacity while\npreserving the visual quality of the stego-image. We also intend to ensure that\nour scheme is resistant to steganalysis attacks.\n This paper proposes a Compressed Sensing Image Steganography (CSIS) scheme to\nachieve our goal while embedding binary data in images. The novelty of our\nscheme is the combination of three components in attaining the above-listed\ngoals. First, we use compressed sensing to sparsify cover image block-wise,\nobtain its linear measurements, and then uniquely select permissible\nmeasurements. Further, before embedding the secret data, we encrypt it using\nthe Data Encryption Standard (DES) algorithm, and finally, we embed two bits of\nencrypted data into each permissible measurement. Second, we propose a novel\ndata extraction technique, which is lossless and completely recovers our secret\ndata. Third, for the reconstruction of the stego-image, we use the least\nabsolute shrinkage and selection operator (LASSO) for the resultant\noptimization problem.\n We perform experiments on several standard grayscale images and a color\nimage, and evaluate embedding capacity, PSNR value, mean SSIM index, NCC\ncoefficients, and entropy. We achieve 1.53 times more embedding capacity as\ncompared to the most recent scheme. We obtain an average of 37.92 dB PSNR\nvalue, and average values close to 1 for both the mean SSIM index and the NCC\ncoefficients, which are considered good. Moreover, the entropy of cover images\nand their corresponding stego-images are nearly the same.\n