보행자 교통안전을 위한 새로운 Boosted 3-D PCA 검출 시스템

Abstract

In this paper, we proposed a new boosted 3-D PCA detection system for pedestrian traffic safety. The proposed system has three steps to improve image detection. At the first step, the proposed algorithm designs a 3-D PCA to solve the performance problem by data balance. At the second step, We design a parallel cross-validation structure to enhance 3-D PCA further. And we also design a modified Adaboost algorithm to improve the detection accuracy performance of 3-D PCA. In order to verify the performance of this system, we experimented with a benchmark over various datasets. As a result, it is better accuracy performance results than PCA-based image detection algorithms.