This paper proposes a face recognition method that is robust against image variations due to arbitrary lighting condition. Though many researches have been carried out on face recognition system, however; there exist some limitations such as illumination, pose, alignment, occlusion, etc. This paper presents a context ontology model making a robust face recognition system on different illumination situations. Our proposed system works on two phases: environmental context ontology building (modelling) and recognition using context ontology. Context ontology is built using data acquisition, context learning and context categorization. The recognition approach is implemented on illumination variant face recognition that takes identified context as input and performs recognition with usual process such as pre-processing, feature extraction, learning, and recognition. We have tested the recognition performance of our proposed model with an international standard FERET face database (our produced synthesized FERET images) and we have achieved a success rate of more than 92%.
展开▼
