Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
4969985 | Pattern Recognition Letters | 2017 | 7 Pages |
Abstract
We propose a guiding network to assist with training a deep convolutional neural network (DCNN) to improve the accuracy of pedestrian detection. The guiding network is adaptively appended to the pedestrian region of the last convolutional layer; the guiding network helps the DCNN to learn the convolutional layers for pedestrian features by focusing on the pedestrian region. The guiding network is used only for training, and therefore does not affect the inference speed. We also explore other factors such as proposal methods and imbalance of training samples. By adopting a guiding network and tackling these factors, our method yields a new state-of-the-art detection accuracy on the Caltech Pedestrian dataset and presents competitive results with the state-of-the-art methods on the INRIA and KITTI datasets.
Related Topics
Physical Sciences and Engineering
Computer Science
Computer Vision and Pattern Recognition
Authors
Sang-Il Jung, Ki-Sang Hong,