Face Detection in Images Using the Cuckoo Optimization Algorithm

Abstract

In recent years, facial recognition has become a prominent area of interest in identity verification systems. Since facial recognition methods rely on computational algorithms designed to identify individuals based on a database of facial images, they have been widely studied. However, less attention has been given to detecting the number of faces within an image and accurately distinguishing facial regions from the background or other objects. Low accuracy and slow processing speed remain two major challenges in the hardware implementation of facial recognition systems. This study investigates face detection in images using the Cuckoo Optimization Algorithm, a metaheuristic approach. By combining image processing techniques with this algorithm, the research aims to determine the number of faces in images and to accurately isolate facial regions. The proposed method was implemented in MATLAB 2015b using the necessary custom commands. The dataset used in this study consisted of 20 images containing varying numbers of faces, collected by the researcher. The results indicate that the Cuckoo Optimization Algorithm can accurately detect the number of faces in an image. Moreover, the algorithm demonstrated high precision in separating facial regions from other areas. It was also observed that increasing the initial population size led to improved optimization performance.                                                                                   

Keywords: Face detection, Metaheuristic algorithm, Cuckoo Optimization.