Detecting the Security Levels of Various Cryptosystems

Due to recent evolutions in multimedia technology, the security of digital data has given rise to serious concerns. To label the imperfection of the current security mechanisms, researchers regularly centralize their efforts on changing the existing protocols. However, throughout the past few decades, several proposed encryption algorithms have been manifest to be insecure, posing a major security risk to sensitive data. It is crucial to choose the best encryption method to defend against these attacks, but the type of data being secured will ascertain which algorithm is best in an especial circumstance. On the other hand, evaluating possible cryptosystems one at a time to find the best substitute can take a while. We offer a method for identifying the security level of picture encryption methods that includes a using of support vector machine to choose appropriate encryption methods (SVM) quickly and precisely. In this study, we also develop a dataset containing traditional encryption security norm, such as entropy, contrast, uniformity, peak signal to noise ratio, mean square error, energy, and correlation. These characteristics are gathered from different cypher images as attributes. Three categories—high, acceptable, and weak—are used to categorize the security quality of dataset labels. We evaluated the performance of our proposed model using several studies (f1-score, recall, precision, and exactness) and the results indicate that this SVM-supported system is successful.