Advancements in Face Recognition through Machine Learning Techniques

Abstract

The majority of disciplines in the modern world rely heavily on face recognition. The identification of fraud and security is one of the most popular disciplines. The process of positioning the facial landmarks on the face to provide precise points for facial recognition is known as facial alignment. Identification and face detection are crucial for detecting fraud. Consequently, the identification of profile and semi-profile facial features is essential for security reasons. The facial-aligned dataset can be used to get the Hourglass model's face alignment, which improves face recognition accuracy when employing the Haar-Cascade Classifier. The precision and accuracy rates are used to gauge performance. When compared to facial recognition using Principal Component Analysis and Support Vector Machines (SVM), it produces better results. SVM, PCA, and deep learning approaches have been utilized in recent years for face recognition; nevertheless, the accuracy rate varies depending on the situation.