Towards a novel biometric system for smart riding club
Résumé
This article presents an investigation on a new Smart Riding Club Biometric System (SRCBS) which recognizes Horse and its rider. Biometric systems have demonstrated an important enhancement in facial recognition from digital images [1]. Facial identity recognition in unconstrained and noncontrolled environment is still a challenging task for human and animal recognition in the wild [2]. In this article, we present a new features extraction technique called Reduced Normalized Gabor Local Binary Pattern (RNGLBP) which is based on the fusion between two basic texture descriptors Gabor and Local Binary Pattern (LBP) used previously in [1]. The Effectiveness of RNGLBP comes from the use of the Reduced Centered Normalization (RCN) operator and the fusion based on the strict application of the Principal Component Analysis (PCA). RNGLBP was developed and tested on three human facial benchmarks Japanese Female Facial Expression (JAFFE), Chon Kanade (CK) and FERET databases and on a new constructed database called Tunisian Horses DataBase of ReGIMLab (THoDBRL'2015). The comprehensive experimental study made has demonstrated that RNGLBP descriptor ranks top the list of several features extraction techniques in both case: human and animal to enhance quality services and to provide more safety for existing riding clubs.