Machine Learning Models for Analyzing Different Cryptosystems' Security Levels

Abstract

Recent developments in multimedia technology have made the security of digital data a vital concern. To address the shortcomings of the current security mechanisms, researchers frequently concentrate their efforts on altering the existing protocols. However, during the past few decades, a number of suggested encryption algorithms have been shown to be unsafe, posing a major security risk to sensitive data. The optimal encryption technique must be used in order to protect against these attacks, but the type of data being secured will determine which algorithm is ideal in a particular circumstance. However, comparing several cryptosystems one at a time to find the best one might take a lot of processing time. We provide a support vector machine-based method for quickly and accurately choosing the appropriate encryption algorithms for photo encryption techniques (SVM). As part of this endeavour, we also generate a dataset using widely used security standards for encryption, including entropy, contrast, homogeneity, peak signal-to-noise ratio, mean square error, energy, and correlation. These factors are utilized as traits that were extracted from different cypher images. Dataset labels are divided into three categories based on their security level: strong, acceptable, and weak. Our recommended model's performance was examined for accuracy, and the findings demonstrate the effectiveness of our SVM-supported system.