Optimization of K-nearest neighbor using particle swarm optimization for face recognition

• Published in: Neural computing & applications Vol. 31; no. 11; pp. 7935 - 7944
• Main Authors: Sasirekha, K., Thangavel, K.
• Format: Journal Article
• Language: English
• Published: London Springer London 01.11.2019; Springer Nature B.V
• Subjects: Algorithms; Ant colony optimization; Artificial Intelligence; Computational Biology/Bioinformatics; Computational Science and Engineering; Computer Science; Data Mining and Knowledge Discovery; Face recognition; Facial recognition technology; Feature extraction; Genetic algorithms; Heuristic methods; Image Processing and Computer Vision; K-nearest neighbors algorithm; Multilayer perceptrons; Optimization; Original Article; Particle swarm optimization; Probability and Statistics in Computer Science; Support vector machines; Webcams; Face biometric; GA; KNN; LBP; ACO; PSO
• ISSN: 0941-0643; 1433-3058
• DOI: 10.1007/s00521-018-3624-9

Face biometric has received more attention to recognize a person in a right way. However, the face recognition is considered to be hard due to size, ethnicity, illumination, pose, various expression, and age. In this work, a novel approach is proposed to recognize the human face based on K-nearest neighbor (KNN) with particle swarm optimization (PSO). Initially, the features are extracted using local binary pattern. The metaheuristic optimization algorithms such as genetic algorithm, PSO, and ant colony optimization are investigated for feature selection. The KNN classifier is optimized using the population-based metaheuristic algorithm PSO. Finally, the face recognition is performed using the proposed PSO–KNN algorithm. In this research, experiments have been conducted on real-time face images collected from 155 subjects each with ten orientations using Logitech Webcam and also on ORL face dataset. The experimental result of the proposed PSO–KNN is compared with other benchmark recognition techniques such as decision table, support vector machine, multilayer perceptron and conventional KNN, to conclude the efficacy of the proposed approach.