Progressive polarization based reflection removal via realistic training data generation

doi:10.1016/j.patcog.2021.108497

CORC > 自动化研究所 > 中国科学院自动化研究所 > 模式识别国家重点实验室 > 三维可视计算

	Progressive polarization based reflection removal via realistic training data generation
	Pang, Youxin 2,3; Yuan, Mengke 2,3; Fu, Qiang 4; Ren, Peiran 1; Yan, Dong-Ming 2,3
刊名	PATTERN RECOGNITION
	2022-04-01
卷号	124 页码:13
关键词	Deep learning Reflection removal Polarization Progressive network Convolutional neural networks
ISSN号	0031-3203
DOI	10.1016/j.patcog.2021.108497
通讯作者	Yuan, Mengke(mengke.yuan@nlpr.ia.ac.cn)
英文摘要	The reflection effect is unavoidable when taking photos through glasses or other transparent materials, which introduces undesired information into pictures. Hence, removing the influence of reflection becomes a key problem in computer vision. One of the main obstacles of recent learning based approaches is the lacking of realistic training data. To address this issue, we introduce a new dataset synthesis method as well as a novel neural network architecture for single image reflection removal. First, we make use of the polarization characteristics of light into the synthesis of datasets, so as to obtain more realistic and diversified training dataset POL. Then, we design a novel Progressive Polarization based Reflection Removal Network ((PR2)-R-2 Net), which preliminary estimates the coarse background layer to guide the final reflection removal. We demonstrate that our method performs better than the state-of-the-art single image reflection removal methods through quantitative and qualitative experimental comparisons. Specifically, the average PSNR of our restored images selected from three representative benchmark datesets: "Real20", "SIR2" and "Nature" is improved at least 0.49 compared with existing methods and reaches to 24.52. (C) 2021 Elsevier Ltd. All rights reserved.
资助项目	National Key R&D Program of China[2019YFB2204104] ; National Natural Science Foundation of China[62102414] ; National Natural Science Foundation of China[62172415] ; National Natural Science Foundation of China[62071157] ; Alibaba Group through Alibaba Innovative Research Program
WOS关键词	SEPARATION
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	ELSEVIER SCI LTD
WOS记录号	WOS:000736972200008
资助机构	National Key R&D Program of China ; National Natural Science Foundation of China ; Alibaba Group through Alibaba Innovative Research Program
内容类型	期刊论文
源URL	[http://ir.ia.ac.cn/handle/173211/47106]
专题	模式识别国家重点实验室_三维可视计算
通讯作者	Yuan, Mengke
作者单位	1.Alibaba Grp, 699 Wang Shang Rd, Hangzhou 310052, Zhejiang, Peoples R China 2.Chinese Acad Sci, Inst Automat, Natl Lab Pattern Recognit, 95 Zhongguancun East Rd, Beijing 100190, PR, Peoples R China 3.Univ Chinese Acad Sci, Sch Artificial Intelligence, Beijing 100049, Peoples R China 4.King Abdullah Univ Sci & Technol, Thuwal 239556900, Saudi Arabia
推荐引用方式 GB/T 7714	Pang, Youxin,Yuan, Mengke,Fu, Qiang,et al. Progressive polarization based reflection removal via realistic training data generation[J]. PATTERN RECOGNITION,2022,124:13.
APA	Pang, Youxin,Yuan, Mengke,Fu, Qiang,Ren, Peiran,&Yan, Dong-Ming.(2022).Progressive polarization based reflection removal via realistic training data generation.PATTERN RECOGNITION,124,13.
MLA	Pang, Youxin,et al."Progressive polarization based reflection removal via realistic training data generation".PATTERN RECOGNITION 124(2022):13.