Frame Prediction Using Recurrent Convolutional Encoder with Residual Learning

DOI

10.2991/amcce-18.2018.77How to use a DOI?

Keywords

residual learning; recurrent convolutional networks; frame prediction.

Abstract

The prediction for the frame of a video is difficult but in urgent need in auto-driving. Conventional methods can only predict some abstract trends of the region of interest. The boom of deep learning makes the prediction for frames possible. In this paper, we propose a novel recurrent convolutional encoder and deconvolutional decoder structure to predict frames. We introduce the residual learning in the convolution encoder structure to solve the gradient issues. The residual learning can transform the gradient backpropagation to an identity mapping. It can reserve the whole gradient information and overcome the gradient issues in Recurrent Neural Networks (RNN) and Convolutional Neural Networks (CNN). Besides, compared with the branches in CNNs and the gated structures in RNNs, the residual learning can save the training time significantly. In the experiments, we use UCF101 dataset to train our networks, the predictions are compared with some state-of-the-art methods. The results show that our networks can predict frames fast and efficiently. Furthermore, our networks are used for the driving video to verify the practicability.

Copyright

© 2018, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Bo-xuan Yue
AU  - Jun Liang
PY  - 2018/05
DA  - 2018/05
TI  - Frame Prediction Using Recurrent Convolutional Encoder with Residual Learning
BT  - Proceedings of the 2018 3rd International Conference on Automation, Mechanical Control and Computational Engineering (AMCCE 2018)
PB  - Atlantis Press
SP  - 449
EP  - 453
SN  - 2352-5401
UR  - https://doi.org/10.2991/amcce-18.2018.77
DO  - 10.2991/amcce-18.2018.77
ID  - Yue2018/05
ER  -