arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 1504.06852cs.CVcs.LG
收藏

FlowNet:用卷积网络学习光流

FlowNet: Learning Optical Flow with Convolutional Networks

University of Freiburg(弗莱堡大学);Technical University of Munich(慕尼黑工业大学)

Philipp Fischer, Alexey Dosovitskiy, Eddy Ilg, Philip Häusser, Caner Hazırbaş, Vladimir Golkov, Patrick van der Smagt, Daniel Cremers, Thomas Brox

摘要:

卷积神经网络(CNN)最近在各种计算机视觉任务中非常成功,特别是在与识别相关的任务中。光流估计并不是CNN成功的任务之一。在本文中,我们构建了适当的CNN,能够解决光流估计问题作为一个监督学习任务。我们提出并比较了两种架构:一个通用的架构和另一个包括一个层,在不同的图像位置的相关特征向量。由于现有的地面实况数据集不足以训练CNN,我们生成一个合成的飞椅数据集。我们发现,在这种不切实际的数据上训练的网络仍然可以很好地推广到现有的数据集,如Sintel和KITTI,在5到10 fps的帧速率下实现了有竞争力的准确性。

英文摘要:

Convolutional neural networks (CNNs) have recently been very successful in a variety of computer vision tasks, especially on those linked to recognition. Optical flow estimation has not been among the tasks where CNNs were successful. In this paper we construct appropriate CNNs which are capable of solving the optical flow estimation problem as a supervised learning task. We propose and compare two architectures: a generic architecture and another one including a layer that correlates feature vectors at different image locations. Since existing ground truth data sets are not sufficiently large to train a CNN, we generate a synthetic Flying Chairs dataset. We show that networks trained on this unrealistic data still generalize very well to existing datasets such as Sintel and KITTI, achieving competitive accuracy at frame rates of 5 to 10 fps.

机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。

↑