A Comparative Computational Efficiency Analysis of LSB and Hybrid Steganographic Methods

Abstract

Electronic Medical Images (EMI) are of various types such as X-rays, CT scans, MRIs, etc. EMIs are generally large, but they vary in dimensions and sizes based on the body part to which they belong. Not much research has been done to secure EMIs due to their varied and large sizes, while ensuring their accessibility for real-time transmission. Where access and storage from third-party storage is performed. Enhancing the security and privacy of EMI while being transmitted or when saved on third-party storage is still a cause of concern. In this research paper, a comparative study is conducted on the Least Significant Bit (LSB) and the Hybrid Method (HM) based on their reduced computational time. In HM, Edge-Based Steganography (EBS) is applied, followed by five layers of cryptography, and separately, LSB steganography is performed to secure the same dataset of 5856 greyscale secret Chest X-ray images, each varying in dimensions and sizes. This paper also discusses EMI and its relevant topics. The HM took an encryption time (ET) of 0.0117 seconds, while the LSB took 5.916 minutes. The decryption time (DT) taken by HM is 0.0142 seconds; on the contrary, the LSB was 13.46 minutes. Concluding based on reduced computational time, making the data security method appropriate for real-time applications (RTA). Thus, ensuring better accessibility to data stored by HM than the LSB method when hiding greyscale Chest X-ray images in the colored cover image. HM performance tests values for Mean square error (MSE) are 0.0000000056, Peak Signal to Noise Ratio (PSNR) is 82.51967, Correlation (R) is 1, Structural Similarity Index Metrics (SSIM) is 1, Signal to Noise Ratio (SNR) is 0.049221, Entropy (E) is 7.8398, and the Number of Pixel Changing Rate (NPCR) is 95.59, which are close to the preferred and good values for Root Mean Square Error (RMSE), Bit Error Rate (BER), etc. Thus, based on computational time and performance tests, these HM values are far superior to the LSB. Therefore, the HM is more practical and can be implemented for real-time applications in the future.