Sharp image estimation from a depth-involved motion-blurred image
文献类型:期刊论文
作者 | Xu, Yuquan; Hu, Xiyuan![]() ![]() |
刊名 | NEUROCOMPUTING
![]() |
出版日期 | 2016 |
卷号 | 171页码:1185-1192 |
关键词 | Image deblurring Depth reconstruction Camera shake Image deconvolution |
英文摘要 | When capturing image of a 3D scene with depth variation, the image blur due to camera shake is no longer spatially invariant but spatially varying which contains the information of the depth of the scene. In this paper, we consider the task of latent image restoration from a single depth-involved motion-blurred (DIMB) image. To easily describe the depth effect, we propose a new framework to model the image blur where the blur kernel is parameterized by the motion curve of the camera and depth value of the scene. Moreover we present an iterative algorithm to alternatively estimate the depth map, motion curve and latent image, based on the minimization of a cost function. To make the estimation of depth map more accurate, a two-stage depth estimation algorithm is also proposed. In the experiment part, the performance of the proposed method is evaluated on many examples including synthetic and real-world cases and shows that our method can produce better results on the DIMB images. (C) 2015 Elsevier B.V. All rights reserved. |
WOS标题词 | Science & Technology ; Technology |
类目[WOS] | Computer Science, Artificial Intelligence |
研究领域[WOS] | Computer Science |
关键词[WOS] | CAMERA |
收录类别 | SCI |
语种 | 英语 |
WOS记录号 | WOS:000364883900117 |
公开日期 | 2016-02-26 |
源URL | [http://ir.ia.ac.cn/handle/173211/10515] ![]() |
专题 | 自动化研究所_智能制造技术与系统研究中心_多维数据分析团队 |
作者单位 | Chinese Acad Sci, Inst Automat, Beijing 100190, Peoples R China |
推荐引用方式 GB/T 7714 | Xu, Yuquan,Hu, Xiyuan,Peng, Silong. Sharp image estimation from a depth-involved motion-blurred image[J]. NEUROCOMPUTING,2016,171:1185-1192. |
APA | Xu, Yuquan,Hu, Xiyuan,&Peng, Silong.(2016).Sharp image estimation from a depth-involved motion-blurred image.NEUROCOMPUTING,171,1185-1192. |
MLA | Xu, Yuquan,et al."Sharp image estimation from a depth-involved motion-blurred image".NEUROCOMPUTING 171(2016):1185-1192. |
入库方式: OAI收割
来源:自动化研究所
浏览0
下载0
收藏0
其他版本
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。