Genetic Algorithm-Based Optimization of SVM-Based Pedestrian Classifier
Ho Gi Jung, Pal Joo Yoon, Jaihie Kim · 2007
Abstract: We developed a pedestrian classifier using GFB(Gabor Filter Bank)-based feature extraction and SVM(Support Vector Machine). Because the SVM uses RBF(Radial Basis Function) and is applied for nonseparable data, learning parameters should be optimized. This paper proposes GA(Ganetic Algorithm)-based optimization of SVM learning parameters. 1.