Coherence and identity learning for arbitrary-length face video generation

Face synthesis is an interesting yet challenging task in computer vision. It is even much harder to generate a portrait video than a single image. In this paper, we propose a novel video generation framework for synthesizing arbitrary-length face videos without any face exemplar or landmark. To over...

Full description

Saved in:
Bibliographic Details
Main Authors: YE, Shuquan, HAN, Chu, LIN, Jiaying, HAN, Guoqiang, HE, Shengfeng
Format: text
Language:English
Published: Institutional Knowledge at Singapore Management University 2021
Subjects:
Online Access:https://ink.library.smu.edu.sg/sis_research/8438
https://ink.library.smu.edu.sg/context/sis_research/article/9441/viewcontent/Ye_2020_facevideo.pdf
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Singapore Management University
Language: English
id sg-smu-ink.sis_research-9441
record_format dspace
spelling sg-smu-ink.sis_research-94412024-01-04T09:56:42Z Coherence and identity learning for arbitrary-length face video generation YE, Shuquan HAN, Chu LIN, Jiaying HAN, Guoqiang HE, Shengfeng Face synthesis is an interesting yet challenging task in computer vision. It is even much harder to generate a portrait video than a single image. In this paper, we propose a novel video generation framework for synthesizing arbitrary-length face videos without any face exemplar or landmark. To overcome the synthesis ambiguity of face video, we propose a divide-and-conquer strategy to separately address the video face synthesis problem from two aspects, face identity synthesis and rearrangement. To this end, we design a cascaded network which contains three components, Identity-aware GAN (IA-GAN), Face Coherence Network, and Interpolation Network. IA-GAN is proposed to synthesize photorealistic faces with the same identity from a set of noises. Face Coherence Network is designed to re-arrange the faces generated by IA-GAN while keeping the inter-frame coherence. Interpolation Network is introduced to eliminate the discontinuity between two adjacent frames and improve the smoothness of the face video. Experimental results demonstrate that our proposed network is able to generate face video with high visual quality while preserving the identity. Statistics show that our method outperforms state-of-the-art unconditional face video generative models in multiple challenging datasets. 2021-01-01T08:00:00Z text application/pdf https://ink.library.smu.edu.sg/sis_research/8438 info:doi/10.1109/ICPR48806.2021.9412380 https://ink.library.smu.edu.sg/context/sis_research/article/9441/viewcontent/Ye_2020_facevideo.pdf http://creativecommons.org/licenses/by-nc-nd/4.0/ Research Collection School Of Computing and Information Systems eng Institutional Knowledge at Singapore Management University Interpolation Cascaded networks Coherence networks Divide and conquer Generative model State of the art Three component Video generation Visual qualities Databases and Information Systems Graphics and Human Computer Interfaces
institution Singapore Management University
building SMU Libraries
continent Asia
country Singapore
Singapore
content_provider SMU Libraries
collection InK@SMU
language English
topic Interpolation
Cascaded networks
Coherence networks
Divide and conquer
Generative model
State of the art
Three component
Video generation
Visual qualities
Databases and Information Systems
Graphics and Human Computer Interfaces
spellingShingle Interpolation
Cascaded networks
Coherence networks
Divide and conquer
Generative model
State of the art
Three component
Video generation
Visual qualities
Databases and Information Systems
Graphics and Human Computer Interfaces
YE, Shuquan
HAN, Chu
LIN, Jiaying
HAN, Guoqiang
HE, Shengfeng
Coherence and identity learning for arbitrary-length face video generation
description Face synthesis is an interesting yet challenging task in computer vision. It is even much harder to generate a portrait video than a single image. In this paper, we propose a novel video generation framework for synthesizing arbitrary-length face videos without any face exemplar or landmark. To overcome the synthesis ambiguity of face video, we propose a divide-and-conquer strategy to separately address the video face synthesis problem from two aspects, face identity synthesis and rearrangement. To this end, we design a cascaded network which contains three components, Identity-aware GAN (IA-GAN), Face Coherence Network, and Interpolation Network. IA-GAN is proposed to synthesize photorealistic faces with the same identity from a set of noises. Face Coherence Network is designed to re-arrange the faces generated by IA-GAN while keeping the inter-frame coherence. Interpolation Network is introduced to eliminate the discontinuity between two adjacent frames and improve the smoothness of the face video. Experimental results demonstrate that our proposed network is able to generate face video with high visual quality while preserving the identity. Statistics show that our method outperforms state-of-the-art unconditional face video generative models in multiple challenging datasets.
format text
author YE, Shuquan
HAN, Chu
LIN, Jiaying
HAN, Guoqiang
HE, Shengfeng
author_facet YE, Shuquan
HAN, Chu
LIN, Jiaying
HAN, Guoqiang
HE, Shengfeng
author_sort YE, Shuquan
title Coherence and identity learning for arbitrary-length face video generation
title_short Coherence and identity learning for arbitrary-length face video generation
title_full Coherence and identity learning for arbitrary-length face video generation
title_fullStr Coherence and identity learning for arbitrary-length face video generation
title_full_unstemmed Coherence and identity learning for arbitrary-length face video generation
title_sort coherence and identity learning for arbitrary-length face video generation
publisher Institutional Knowledge at Singapore Management University
publishDate 2021
url https://ink.library.smu.edu.sg/sis_research/8438
https://ink.library.smu.edu.sg/context/sis_research/article/9441/viewcontent/Ye_2020_facevideo.pdf
_version_ 1787590749518299136