Real-time Pedestrian Detection Using LIDAR and Convolutional Neural Networks.

This paper presents a pedestrian detection system based on sensor fusion of LIDAR and convolutional neural network based image classifier. This method achieves over 10 frames/second processing speed by utilizing the LIDAR. The focus of this paper is the evaluation of the effect of fusing the LIDAR compared to the image-only system. The evaluation results indicate that fusing the LIDAR can reduce the number of false positives by a factor of2 and reduce the processing time by a factor of 4. The single frame detection accuracy of the system is above 90% when there is 1 false positive/second. For the covering abstract see E134653.