The Optimal pedestrian detection algorithm based on dynamic adaptive region convolution model

摘要

The traditional target detection and identification algorithm is difficult to adapt to the massive data. And the expression of the method it relies on is designed by means of manual, which is not only very time-consuming, but also very dependence for professional knowledge and data itself. Aiming at the problem of pedestrian detection in complex environment, a pedestrian detection algorithm based on dynamic adaptive region convolution model is proposed. The detection results on INRIA pedestrian data set show the improved detection performance. And the proposed method can detect pedestrians successfully in most complex background.