Feature Selection Based on Machine Learning Algorithms: A weighted Score Feature Importance Approach for Facial Authentication

The emergence of biometric-based face detection and recognition technology has attracted the attention of all industries. Facial recognition allows you to recognize people by their facial features. This technology has been a part of our daily lives and has been utilized in security, forensic investigation, and check-ins at airports. Feature selection allows for finding the most distinct facial biometric traits from the first feature set. In addition to identifying the most important features, feature selection also aids in lowering the feature size dimension. It is an essential step in the process since biometric authentication occurs in real-time. This paper presents a novel feature selection method based on the fusion feature importance scores of machine learning models. A weighted score level fusion method based on the Genetic Algorithm (GA) is used to combine the scores collected from various machine learning models. The objective function, which represents the accuracy of the biometrics system, is optimized by the GA, which is utilized as an optimizer to choose the optimal weights. The Fetch Olivetti Faces database has been used to test the proposed method. According to the stated results, the proposed approach increased accuracy from 93.5 percent to 95.62 percent while enabling a decrease in feature size from 4096 to 964.