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 1903.00271cs.CVcs.LG
收藏

用于视频预测的频域Transformer网络

Frequency Domain Transformer Networks for Video Prediction

University of Bonn(波恩大学)

Hafez Farazi, Sven Behnke

摘要:

视频预测的任务是在给定前一帧的情况下预测下一帧。尽管最近取得了很大进展,但这项任务仍然具有挑战性,主要是由于空间域中的高非线性。为了解决这个问题,我们提出了一种新的架构,频域Transformer网络(FDTN),这是一个端到端的可学习模型,估计和使用的信号在频域中的变换。实验结果表明,该方法的性能优于一些广泛使用的视频预测方法,如视频梯形网络(VLN)和预测门控金字塔(PGP)。

英文摘要:

The task of video prediction is forecasting the next frames given some previous frames. Despite much recent progress, this task is still challenging mainly due to high nonlinearity in the spatial domain. To address this issue, we propose a novel architecture, Frequency Domain Transformer Network (FDTN), which is an end-to-end learnable model that estimates and uses the transformations of the signal in the frequency domain. Experimental evaluations show that this approach can outperform some widely used video prediction methods like Video Ladder Network (VLN) and Predictive Gated Pyramids (PGP).

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。

↑