Improving the Quality of the Face Recognition Using LBPH in Machine Learning

Abstract

The objective of this article is to recognize facial features from images using face recognition approaches. For preprocessed face images with equalized histograms, the suggested solution employs a local binary pattern histogram (LBPH). One of the long-standing issues with computer vision is accurate face detection and recognition. The Local Binary Pattern (LBP) is a better facial descriptor in face recognition, in recent study. A person's identity, sentiments, and ideas may be more easily discernible from their face. Everyone wants to feel safe from unauthorized authentication in the current world. To improve security, face detection and recognition have joined the scene and are tackling the most challenging challenge of effectively recognizing faces without creating any false identities. The histogram values are extracted and joined into a single vector. After applying these methods, the training loss decreases and the validation of accuracy rise by over 96.5%. This vector compares the facial likenesses and produces the most advantageous outcome.