FDFlowNet: Fast Optical Flow Estimation using a Deep Lightweight Network
This work addresses the need for real-time optical flow estimation in applications like robotics and autonomous driving, though it is incremental as it builds on existing deep learning methods.
The authors tackled the problem of high computational complexity in deep optical flow estimation by introducing FDFlowNet, a lightweight network that achieves similar or better accuracy on KITTI and Sintel benchmarks while being about 2 times faster than PWC-Net.
Significant progress has been made for estimating optical flow using deep neural networks. Advanced deep models achieve accurate flow estimation often with a considerable computation complexity and time-consuming training processes. In this work, we present a lightweight yet effective model for real-time optical flow estimation, termed FDFlowNet (fast deep flownet). We achieve better or similar accuracy on the challenging KITTI and Sintel benchmarks while being about 2 times faster than PWC-Net. This is achieved by a carefully-designed structure and newly proposed components. We first introduce an U-shape network for constructing multi-scale feature which benefits upper levels with global receptive field compared with pyramid network. In each scale, a partial fully connected structure with dilated convolution is proposed for flow estimation that obtains a good balance among speed, accuracy and number of parameters compared with sequential connected and dense connected structures. Experiments demonstrate that our model achieves state-of-the-art performance while being fast and lightweight.