Face Recognition Based on LBP Multi-feature Extraction Fusion and KNN

Abstract

The widespread application of facial recognition technology has enriched people's digital lives. In such a digital age, data intelligence based on cognitive computing has transformed an individual from a statistically average person into an object that can be analyzed for data independently. This change heralds the arrival of a data analysis society. The facial recognition system integrates multiple professional technologies such as artificial intelligence, machine recognition, machine learning, model theory, expert systems, and video image processing. At the same time, it needs to combine the theory and implementation of intermediate value processing, and is the latest application of biometric recognition. The realization of its core technology demonstrates the transformation from weak artificial intelligence to strong artificial intelligence. To effectively extract facial features from different angles and postures under complex lighting conditions, this paper proposes a novel LBP multi-feature extraction fusion and KNN method. This paper proposes two improved local binary patterns (LBP), namely weighted fusion of triangular features with local binary pattern (WLBP) and Tulapuras local binary pattern (TLBP). They are fused with the Gabor filter and then classified by using the KNN classifier. The experiment adopts the YALEB, ORL and FERET face databases to conduct experiments. The experimental results show that the proposed method significantly improves the accuracy of face recognition under complex lighting and multi-angle and pose conditions compared with the traditional single LBP method.