An intelligent system for detecting and recognizing human faces

Abstract

Detecting human faces in any image remains a challenging problem in computer vision. The use of artificial intelligence and deep learning techniques has significantly enhanced performance in the areas of face detection and analysis. The progress is supported by both standard and complex datasets, which enable models to be trained across various scenarios. The advancement of face detection technologies is being utilized across various fields, including security systems, marketing, and healthcare systems, with innovations to enhance the system’s speed and accuracy. Local database (LDB) models were used for the two models: a training set that employs a convolutional neural network (CNN) and an accompanying test set which assesses its performance using a confusion matrix (CM), and three image sizes (32,64 and 128). The findings indicate that model performance improved with the use of local data and across differing image sizes, with the pinnacle being an image size of 64 with the use of an adaptive network layer, achieving a result with an accuracy of: Accuracy (ACC) = 0.94375, Precision (P) = 0.90149, True Positive Rate (TPR) = 0.8875.