A novel pedestrian detection method based on Cost-Sensitive Support Vector Machine and Chaotic Particle Swarm Optimization with T mutation

This paper presents a novel pedestrian detection method based on chaotic particle swarm optimization with T mutation (CTPSO) and cost-sensitive support vector machine (CS-SVM). In order to solve the problem of class-imbalanced in pedestrian detection, a new improve SVM named CS-SVM is proposed, which is based on the idea of assigning different weights to the errors of the two classes when the numbers of data samples from each class are imbalanced. In addition, a new type of PSO called CTPSO is used to select suitable parameters of CS-SVM, which could improve the classification ability of CS-SVM prominently. CTPSO is a novel optimization algorithm, which not only has strong global search capability but also helps to find the optimum quickly by using chaos queues and T mutation. The experiment carried out on videos from INRIA, MIT and Daimler datasets, result indicates that the effectiveness and efficiency of the proposed method, which can achieve higher accuracy than other three state of the art algorithms.