Nonlinear camera response functions and image deblurring: Theoretical analysis and practice

Yu Wing Tai, Xiaogang Chen, Sunyeong Kim, Seon Joo Kim, Feng Li, Jie Yang, Jingyi Yu, Yasuyuki Matsushita, Michael S. Brown

Research output: Contribution to journalArticle

47 Citations (Scopus)

Abstract

This paper investigates the role that nonlinear camera response functions (CRFs) have on image deblurring. We present a comprehensive study to analyze the effects of CRFs on motion deblurring. In particular, we show how nonlinear CRFs can cause a spatially invariant blur to behave as a spatially varying blur. We prove that such nonlinearity can cause large errors around edges when directly applying deconvolution to a motion blurred image without CRF correction. These errors are inevitable even with a known point spread function (PSF) and with state-of-the-art regularization-based deconvolution algorithms. In addition, we show how CRFs can adversely affect PSF estimation algorithms in the case of blind deconvolution. To help counter these effects, we introduce two methods to estimate the CRF directly from one or more blurred images when the PSF is known or unknown. Our experimental results on synthetic and real images validate our analysis and demonstrate the robustness and accuracy of our approaches.

Original languageEnglish
Article number6461888
Pages (from-to)2498-2512
Number of pages15
JournalIEEE Transactions on Pattern Analysis and Machine Intelligence
Volume35
Issue number10
DOIs
Publication statusPublished - 2013 Sep 3

Fingerprint

Image Deblurring
Response Function
Theoretical Analysis
Camera
Cameras
Optical transfer function
Deconvolution
Blind Deconvolution
Deblurring
Function Estimation
Motion
Estimation Algorithms
Regularization
Nonlinearity
Robustness
Unknown
Invariant
Experimental Results
Estimate
Demonstrate

All Science Journal Classification (ASJC) codes

  • Software
  • Computer Vision and Pattern Recognition
  • Computational Theory and Mathematics
  • Artificial Intelligence
  • Applied Mathematics

Cite this

Tai, Yu Wing ; Chen, Xiaogang ; Kim, Sunyeong ; Kim, Seon Joo ; Li, Feng ; Yang, Jie ; Yu, Jingyi ; Matsushita, Yasuyuki ; Brown, Michael S. / Nonlinear camera response functions and image deblurring : Theoretical analysis and practice. In: IEEE Transactions on Pattern Analysis and Machine Intelligence. 2013 ; Vol. 35, No. 10. pp. 2498-2512.
@article{68dca0d4ff3d45ebb8a39fd12689a305,
title = "Nonlinear camera response functions and image deblurring: Theoretical analysis and practice",
abstract = "This paper investigates the role that nonlinear camera response functions (CRFs) have on image deblurring. We present a comprehensive study to analyze the effects of CRFs on motion deblurring. In particular, we show how nonlinear CRFs can cause a spatially invariant blur to behave as a spatially varying blur. We prove that such nonlinearity can cause large errors around edges when directly applying deconvolution to a motion blurred image without CRF correction. These errors are inevitable even with a known point spread function (PSF) and with state-of-the-art regularization-based deconvolution algorithms. In addition, we show how CRFs can adversely affect PSF estimation algorithms in the case of blind deconvolution. To help counter these effects, we introduce two methods to estimate the CRF directly from one or more blurred images when the PSF is known or unknown. Our experimental results on synthetic and real images validate our analysis and demonstrate the robustness and accuracy of our approaches.",
author = "Tai, {Yu Wing} and Xiaogang Chen and Sunyeong Kim and Kim, {Seon Joo} and Feng Li and Jie Yang and Jingyi Yu and Yasuyuki Matsushita and Brown, {Michael S.}",
year = "2013",
month = "9",
day = "3",
doi = "10.1109/TPAMI.2013.40",
language = "English",
volume = "35",
pages = "2498--2512",
journal = "IEEE Transactions on Pattern Analysis and Machine Intelligence",
issn = "0162-8828",
publisher = "IEEE Computer Society",
number = "10",

}

Nonlinear camera response functions and image deblurring : Theoretical analysis and practice. / Tai, Yu Wing; Chen, Xiaogang; Kim, Sunyeong; Kim, Seon Joo; Li, Feng; Yang, Jie; Yu, Jingyi; Matsushita, Yasuyuki; Brown, Michael S.

In: IEEE Transactions on Pattern Analysis and Machine Intelligence, Vol. 35, No. 10, 6461888, 03.09.2013, p. 2498-2512.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Nonlinear camera response functions and image deblurring

T2 - Theoretical analysis and practice

AU - Tai, Yu Wing

AU - Chen, Xiaogang

AU - Kim, Sunyeong

AU - Kim, Seon Joo

AU - Li, Feng

AU - Yang, Jie

AU - Yu, Jingyi

AU - Matsushita, Yasuyuki

AU - Brown, Michael S.

PY - 2013/9/3

Y1 - 2013/9/3

N2 - This paper investigates the role that nonlinear camera response functions (CRFs) have on image deblurring. We present a comprehensive study to analyze the effects of CRFs on motion deblurring. In particular, we show how nonlinear CRFs can cause a spatially invariant blur to behave as a spatially varying blur. We prove that such nonlinearity can cause large errors around edges when directly applying deconvolution to a motion blurred image without CRF correction. These errors are inevitable even with a known point spread function (PSF) and with state-of-the-art regularization-based deconvolution algorithms. In addition, we show how CRFs can adversely affect PSF estimation algorithms in the case of blind deconvolution. To help counter these effects, we introduce two methods to estimate the CRF directly from one or more blurred images when the PSF is known or unknown. Our experimental results on synthetic and real images validate our analysis and demonstrate the robustness and accuracy of our approaches.

AB - This paper investigates the role that nonlinear camera response functions (CRFs) have on image deblurring. We present a comprehensive study to analyze the effects of CRFs on motion deblurring. In particular, we show how nonlinear CRFs can cause a spatially invariant blur to behave as a spatially varying blur. We prove that such nonlinearity can cause large errors around edges when directly applying deconvolution to a motion blurred image without CRF correction. These errors are inevitable even with a known point spread function (PSF) and with state-of-the-art regularization-based deconvolution algorithms. In addition, we show how CRFs can adversely affect PSF estimation algorithms in the case of blind deconvolution. To help counter these effects, we introduce two methods to estimate the CRF directly from one or more blurred images when the PSF is known or unknown. Our experimental results on synthetic and real images validate our analysis and demonstrate the robustness and accuracy of our approaches.

UR - http://www.scopus.com/inward/record.url?scp=84883144504&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84883144504&partnerID=8YFLogxK

U2 - 10.1109/TPAMI.2013.40

DO - 10.1109/TPAMI.2013.40

M3 - Article

C2 - 23969392

AN - SCOPUS:84883144504

VL - 35

SP - 2498

EP - 2512

JO - IEEE Transactions on Pattern Analysis and Machine Intelligence

JF - IEEE Transactions on Pattern Analysis and Machine Intelligence

SN - 0162-8828

IS - 10

M1 - 6461888

ER -