Kartika Candra Kirana, Slamet Wibawanto, Nur Hidayah, Gigih Prasetyo Cahyono
In previous studies, we have developed a face detection system. Even though, we have not been able to overcome the face searching. In addition, there are many possibilities for various facial positions. The sliding window technique that gained popularity as well-known techniques. It searches the face model in the entire image quickly by evaluating the information heuristic each window. In order to prevent the dissipation exploring of the non-face candidate, we propose an ant system which can choose routes based on heuristic information to optimize the sliding window technique in face detection. Ant colony algorithm and the Viola-Jones algorithm achieve an iteration average of up to 14150 and 19200. In addition, they both reach computational time up to 50.699 FPS and 50.358 FPS. Although the results are almost the same and the test data is limited, our method is worthy of being excellence. © 2019 IEEE.
Department of Electrical Engineering, Universitas Negeri Malang, Malang, Indonesia; Department of Guidance and Counseling, Universitas Negeri Malang, Malang, Indonesia; Software Engineering, Visionet Data International, Malang, Indonesia
