arrow
返回

High-Quality Video Generation from Static Structural Annotations

delete2020-05-28
delete3
PRE
AI
L
Lu Sheng *
J
Junting Pan
J
Jiaming Guo
J
Jing Shao
C
Chen Change Loy
DOI:10.1007/s11263-020-01334-xdelete
delete原文链接
delete原文求助
delete分享
delete收藏
摘要

摘要

En 中文
This paper proposes a novel unsupervised video generation that is conditioned on a single structural annotation map, which in contrast to prior conditioned video generation approaches, provides a good balance between motion flexibility and visual quality in the generation process. Different from end-to-end approaches that model the scene appearance and dynamics in a single shot, we try to decompose this difficult task into two easier sub-tasks in a divide-and-conquer fashion, thus achieving remarkable results overall. The first sub-task is an image-to-image (I2I) translation task that synthesizes high-quality starting frame from the input structural annotation map. The second image-to-video (I2V) generation task applies the synthesized starting frame and the associated structural annotation map to animate the scene dynamics for the generation of a photorealistic and temporally coherent video. We employ a cycle-consistent flow-based conditioned variational autoencoder to capture the long-term motion distributions, by which the learned bi-directional flows ensure the physical reliability of the predicted motions and provide explicit occlusion handling in a principled manner. Integrating structural annotations into the flow prediction also improves the structural awareness in the I2V generation process. Quantitative and qualitative evaluations over the autonomous driving and human action datasets demonstrate the effectiveness of the proposed approach over the state-of-the-art methods. The code has been released:.
Keyword:
Unsupervised learning
Conditioned generative model
Image and video synthesis
Motion prediction and estimatiovn
AI总结

AI总结

对已上传原文的论文进行重点信息的提取,主要内容包括:简要概述、研究摘要、背景介绍、关键亮点、图文解析、展望与总结。

期刊

International Journal of Computer Vision 封面图
International Journal of Computer Vision
IF:
9.3
论文数:
3.9K
被引数:
2.8W

机构

B
Beihang University
学者数:
5.2W
论文数: 4.1W
被引数: 37
N
Nanyang Technological University
学者数:
4.9W
论文数: 4.8W
被引数: 8.1W
C
Chinese University of Hong Kong
学者数:
3.4W
论文数: 3.2W
被引数: 5.6W
I
institute of computing technology, cas
学者数:
1.0K
论文数: 878
被引数: 1
C
chinese academy of sciences
学者数:
56.7W
论文数: 45.0W
被引数: 704
学者 查看更多机构
引用论文

引用论文

err分享
err收藏
A segmentation approach for the reproducible extraction and quantification of knickpoints from river long profiles
err2019-02-18
err0
errOAAI
errBoris Gailleton; Simon M. Mudd; Fiona J. Clubb; Daniel Peifer; Martin D. Hurst
err分享
err收藏
Formation of new C–O and C–N bonds via base promoted Csp2–Csp3 bond cleavage of α-nitro ketone
err2015-12-01
err0
PREAI
errManas Jyoti Sarma; Arun Jyoti Borah; Kamal Krishna Rajbongshi; Prodeep Phukan
err分享
err收藏
err分享
err收藏
Deep Reinforcement Learning: A Brief Survey深度强化学习: 简要综述
err2017-11-01
err2.4K
errOAAI
errArulkumaran, Kai; Deisenroth, Marc Peter; Brundage, Miles; Bharath, Anil Anthony
err分享
err收藏
Use of a soil moisture network for drought monitoring in the Czech Republic
err2011-06-09
err0
PREAI
errMartin Mozny; Mirek Trnka; Zdenek Zalud; Petr Hlavinka; Jiri Nekovar; Vera Potop; Michal Virag
err分享
err收藏
学者 查看更多内容