Synthetic datasets play a critical role in pre-training CNN models for optical flow, but they are painstaking to generate and hard to adapt to new applications. To automate the process, we present AutoFlow, a simple and effective method to render training data for optical flow that optimizes the performance of a model on a target dataset. AutoFlow takes a layered approach to render synthetic data, where the motion, shape, and appearance of each layer are controlled by learnable hyperparameters. Experimental results show that AutoFlow achieves state-of-the-art accuracy in pre-training both PWC-Net and RAFT. Our code and data are available at https://autoflow-google.github.io .
@article{arxiv.2104.14544,
title = {AutoFlow: Learning a Better Training Set for Optical Flow},
author = {Deqing Sun and Daniel Vlasic and Charles Herrmann and Varun Jampani and Michael Krainin and Huiwen Chang and Ramin Zabih and William T. Freeman and Ce Liu},
journal= {arXiv preprint arXiv:2104.14544},
year = {2021}
}