01-05-2017, 03:47 PM
Feature selection (FS) is a global optimization problem in automatic learning that reduces the number of features, eliminates irrelevant, noisy and redundant data and results in acceptable recognition accuracy. It is the most important step that affects the performance of a pattern recognition system. This paper presents a new feature selection algorithm based on particle swarm optimization (PSO). The algorithm is applied to the coefficients extracted by two characteristic extraction techniques: the discrete cosine transform (DCT) and the discrete wavelet transform (DWT).
The proposed PSO-based feature selection algorithm is used to find the feature space for the optimal subset of features where features are carefully selected according to a well-defined discrimination criterion. Evolution is driven by a fitness function defined in terms of maximizing class separation (dispersion index). The performance of the classifier and the length of the selected feature vector are considered for performance evaluation using the ORL face database. Experimental results show that the PSO-based feature selection algorithm was found to generate excellent recognition results with the minimum set of selected characteristics.
Face recognition (FR) has emerged as one of the most widely researched research topics covering multiple disciplines such as pattern recognition, signal processing and computer vision. This is due to its numerous important applications in identity authentication, access control to security, intelligent interaction between humans and computers and the automatic indexing of databases of images and videos. Many approaches have been developed to deal with surveys; An excellent survey document on different facial recognition techniques can be found. The success of any FR methodology depends to a large extent on the particular choice of characteristics used by the classifier (pattern). It is known that a good feature extractor for a facial recognition system is claimed to select as best as possible the best discriminating characteristics that are not sensitive to arbitrary environmental variations such as variations in pose, scale, illumination and facial expressions. Characteristic extraction algorithms are mainly divided into two categories: extraction of geometric characteristics and extraction of statistical (algebraic) characteristics.
Particle Swarm Optimization (PSO) is a population-based stochastic optimization technique originally proposed by James Kennedy and Russell C. Eberhart in 1995. PSO is a search algorithm based on the simulation of the behavior of birds within a herd . Definitions of various technical terms commonly used in PSO can be found in. The swarm is a population of particles. Each particle represents a potential solution to the problem being solved. The best personal (pbest) of a given particle is the position of the particle that has provided the greatest success (ie, the maximum value given by the classification method used). The best local (lbest) is the position of the best particle member in the neighbourhood of a given particle. The best overall (gbest) is the position of the best particle of the whole swarm. The leader is the particle that is used to guide another particle towards better regions of the search space.
The proposed PSO-based feature selection algorithm is used to find the feature space for the optimal subset of features where features are carefully selected according to a well-defined discrimination criterion. Evolution is driven by a fitness function defined in terms of maximizing class separation (dispersion index). The performance of the classifier and the length of the selected feature vector are considered for performance evaluation using the ORL face database. Experimental results show that the PSO-based feature selection algorithm was found to generate excellent recognition results with the minimum set of selected characteristics.
Face recognition (FR) has emerged as one of the most widely researched research topics covering multiple disciplines such as pattern recognition, signal processing and computer vision. This is due to its numerous important applications in identity authentication, access control to security, intelligent interaction between humans and computers and the automatic indexing of databases of images and videos. Many approaches have been developed to deal with surveys; An excellent survey document on different facial recognition techniques can be found. The success of any FR methodology depends to a large extent on the particular choice of characteristics used by the classifier (pattern). It is known that a good feature extractor for a facial recognition system is claimed to select as best as possible the best discriminating characteristics that are not sensitive to arbitrary environmental variations such as variations in pose, scale, illumination and facial expressions. Characteristic extraction algorithms are mainly divided into two categories: extraction of geometric characteristics and extraction of statistical (algebraic) characteristics.
Particle Swarm Optimization (PSO) is a population-based stochastic optimization technique originally proposed by James Kennedy and Russell C. Eberhart in 1995. PSO is a search algorithm based on the simulation of the behavior of birds within a herd . Definitions of various technical terms commonly used in PSO can be found in. The swarm is a population of particles. Each particle represents a potential solution to the problem being solved. The best personal (pbest) of a given particle is the position of the particle that has provided the greatest success (ie, the maximum value given by the classification method used). The best local (lbest) is the position of the best particle member in the neighbourhood of a given particle. The best overall (gbest) is the position of the best particle of the whole swarm. The leader is the particle that is used to guide another particle towards better regions of the search space.
Velocity is the vector that determines the direction in which a particle needs to "fly" (move), to improve its current position. The weight of inertia, denoted by W, is used to control the impact of the previous history of velocities on the current velocity of a given particle. The learning factor represents the attraction that a particle has for its own success (C1 - cognitive learning factor) or that of its neighbors (C2 - social learning factor). Both C1 and C2 are normally defined as constants. Finally, the neighborhood topology determines the set of particles that contribute to the calculation of the highest value of a given particle.