Facial Landmark Detection using Affine Graph Matching and a Genetic Search Algorithm


  • Resmana Lim Faculty of Industrial Technology, Petra Christian University




facial landmark detection, face recognition, affine graph matching, genetic algorithm, Gabor filtering.


This paper proposes a method that finds landmark points on the face, which is one of the main tasks in a face recognition system. Salient facial landmark detection is important because it enables face normalization and leads to size and orientation invariant face recognition. The presented approach is based on an affine graph matching technique and uses a genetic algorithm to perform the search. The feasibility of our methodology for detection tasks related to face landmark point detection has been deployed using the ORL face image database. Experiments show satisfactory results under relatively wide conditions. The GA searching approach is essential because it effectively searches the solution space.. Abstract in Bahasa Indonesia : Paper ini membicarakan sebuah metode pencarian titik-titik penting pada citra wajah, yang mana ini adalah langkah penting dalam system pengenalan wajah. Titik-titik penting wajah (seperti mata, hidung dan mulut) adalah sangat penting untuk proses normalisasi wajah sehingga dihasilkan system pengenalan wajah yang handal terhadap variasi ukuran dan orientasi. Pendekatan yang digunakan disini menggunakan teknik affine graph matching dan algoritma genetika untuk pencarian titik-titik penting wajah. Sistem diuji menggunakan database wajah dari ORL. Percobaan menunjukan hasil yang cukup menggembirakan. Pencarian dengan algoritma genetika sangat penting disini sebab dapat mencari solusi secara efektif. Kata kunci : deteksi titi-titik penting wajah, pengenalan wajah, affine graph matching, algoritma genetika, Gabor filter.