Multi-scale Pedestrian Detection by Use of AdaBoost Learning Algorithm
Wei Guo, Ya Xiao, Guodong Zhang · 2014
Pedestrian detection has a wide range of applications in visual surveillance, driver assistance systems. It is also very important in computer vision and pattern recognition. In our study, we proposed a multi-scale scheme for pedestrian detection. The scheme of pedestrian detection consisted of two steps for construction a strong classifier and multi-scale detection. The strong classifier, a collection of weak classifiers, was built by use of AdaBoost learning algorithm based on the Harr-like features. Then, the strong classifier was employed to detect pedestrians in the multi-scale images, and the detection results were merged. In our experiment, the proposed multi-scale detection scheme reported 0.35 false positives per image at the sensitivity to 89.3%. This indicates that the multi-scale scheme for pedestrian detection achieves a high performance.