A Comparison of ANFIS, SVM and KNN models for the face recognition in unconstrained environment
Hossam Fraihat · International Journal of Advanced Trends in Computer Science and Engineering · 2020
For the rapid extending applications of face recognition as an important technique for identification over other biometric features, different types of databases of images have been generated and published.The existing face datasets can be collected either in controlled environments such that the ones used for drivers' license and passport or unconstrained environments such that with variation in size, position, pose, lighting, expression, background, camera quality, occlusion, age, and gender.Many techniques were proposed to identify different acquired characters, resulting in wide various accuracies and other measures.Supervised machine learning techniques especially Adaptive-network-based fuzzy inference (ANFIS), K-nearest neighbor (KNN|) and support vector machine (SVM) have been widely used as models in this context.In this paper, we study the accuracy result of those models when used in various types of database, and conclude with the best conditions for satisfactory recognition results.