Removing Motion Blur using Natural Image Statistics

Johannes Herwig, Timm Linder, Josef Pauli

2014

Abstract

We tackle deconvolution of motion blur in hand-held consumer photography with a Bayesian framework combining sparse gradient and color priors for regularization. We develop a closed-form optimization utilizing iterated re-weighted least squares (IRLS) with a Gaussian approximation of the regularization priors. The model parameters of the priors can be learned from a set of natural images which resemble common image statistics. We throughly evaluate and discuss the effect of different regularization factors and make suggestions for reasonable values. Both gradient and color priors are current state-of-the-art. In natural images the magnitude of gradients resembles a kurtotic hyper-Laplacian distribution, and the two-color model exploits the observation that locally any color is a linear approximation between some primary and secondary colors. Our contribution is integrating both priors into a single optimization framework and providing a more detailed derivation of their optimization functions. Our re-implementation reveals different model parameters than previously published, and the effectiveness of the color priors alone are explicitly examined. Finally, we propose a context-adaptive parameterization of the regularization factors in order to avoid over-smoothing the deconvolution result within highly textured areas.

References

  1. Cecchi, G. A., Rao, A. R., Xiao, Y., and Kaplan, E. (2010). Statistics of natural scenes and cortical color processing. Journal of Vision, 10(11):1-13.
  2. Chan, T. and Shen, J. (2005). Image Processing And Analysis: Variational, Pde, Wavelet, And Stochastic Methods. Siam.
  3. Cho, S., Wang, J., and Lee, S. (2012). Video deblurring for hand-held cameras using patch-based synthesis. ACM Trans. on Graphics, 31(4):64:1-64:9.
  4. Cui, G., Feng, H., Xu, Z., Li, Q., and Chen, Y. (2014). A modified Richardson-Lucy algorithm for a single image with adaptive reference maps. Optics & Laser Technology, 58:100-109.
  5. Dong, W., Feng, H., Xu, Z., and Li, Q. (2012a). Blind image deconvolution using the fields of experts prior. Optics Communications, 285:5051-5061.
  6. Dong, W., Shi, G., Li, X., Zhang, L., and Wu, X. (2012b). Image reconstruction with locally adaptive sparsity and nonlocal robust regularization. Image Communication, 27:1109-1122.
  7. Faraway, J. J. (2002). Practical regression and ANOVA using R. Online. http://cran.r-project.org/ doc/contrib/Faraway-PRA.pdf.
  8. Gentle, J. (2007). Matrix algebra: theory, computations, and applications in statistics. Springer texts in statistics. Springer, New York, NY.
  9. Gupta, A., Joshi, N., Zitnick, L., Cohen, M., and Curless, B. (2010). Single image deblurring using motion density functions. In ECCV 7810: Proc. of the 10th Europ. Conf. on Comp. Vision.
  10. Huang, J. (2000). Statistics of natural images and models. PhD thesis, Brown Univ., Providence.
  11. Ito, A., Sankaranarayanan, A. C., Veeraraghavan, A., and Baraniuk, R. G. (2013). Blurburst: Removing blur due to camera shake using multiple images. ACM Trans. Graph. Submitted.
  12. Ji, H. and Wang, K. (2012). A two-stage approach to blind spatially-varying motion deblurring. In IEEE Conf. on Comp. Vision and Pattern Recogn. (CVPR), pages 73-80. IEEE Comp. Soc.
  13. Joshi, N., Zitnick, C., Szeliski, R., and Kriegman, D. (2009). Image deblurring and denoising using color priors. In IEEE Conf. on Comp. Vision and Pattern Recognition (CVPR), pages 1550-1557. IEEE Comp. Soc.
  14. Krishnan, D. and Fergus, R. (2009). Fast image deconvolution using hyper-laplacian priors. In Neural Inform. Proc. Sys.
  15. Levin, A., Fergus, R., Durand, F., and Freeman, W. T. (2007a). Deconvolution using natural image priors. Technical report, MIT.
  16. Levin, A., Fergus, R., Durand, F., and Freeman, W. T. (2007b). Image and depth from a conventional camera with a coded aperture. ACM Trans. Graph., 26.
  17. Levin, A. and Weiss, Y. (2007). User assisted separation of reflections from a single image using a sparsity prior. IEEE Trans. Pattern Anal. and Mach. Intell., 29(9):1647-1654.
  18. Li, X., Pan, J., Lin, Y., and Su, Z. (2013). Fast blind deblurring via normalized sparsity prior. Journal of Inform. & Comp. Sc., 10(16):5083-5091.
  19. Lin, H. T., Tai, Y.-W., and Brown, M. S. (2011). Motion regularization for matting motion blurred objects. IEEE Trans. Pattern Anal. and Mach. Intell., 33(11):2329- 2336.
  20. Liu, R., Li, Z., and Jia, J. (2008). Image partial blur detection and classification. In IEEE Conf. on Comp. Vision and Pattern Recognition (CVPR), pages 1-8. IEEE Comp. Soc.
  21. Marichal, X., Ma, W.-Y., and Zhang, H.-J. (1999). Blur determination in the compressed domain using DCT information. In Intern. Conf. on Image Proc. (ICIP), pages 386-390.
  22. Meer, P. (2004). Robust techniques for computer vision. In Medioni, G. and Kang, S. B., editors, Emerging Topics in Computer Vision, chapter 4. Prentice Hall PTR, Upper Saddle River, NJ.
  23. Scales, J. A. and Gersztenkorn, A. (1988). Robust methods in inverse theory. Inverse Problems, 4(4):1071.
  24. Scales, J. A., Gersztenkorn, A., and Treitel, S. (1988). Fast lp solution of large, sparse, linear systems: Application to seismic travel time tomography. Journal of Computational Physics, 75:314-333.
  25. Schmidt, U., Rother, C., Nowozin, S., Jancsary, J., and Roth, S. (2013). Discriminative non-blind deblurring. In IEEE Conf. on Comp. Vision and Pattern Recognition (CVPR), pages 604-611. IEEE Comp. Soc.
  26. Schmidt, U., Schelten, K., and Roth, S. (2011). Bayesian deblurring with integrated noise estimation. In IEEE Conf. on Comp. Vision and Pattern Recognition (CVPR), pages 2625-2632. IEEE Comp. Soc.
  27. Shi, M.-z., Xu, T.-f., Feng, L., Liang, J., and Zhang, K. (2013). Single image deblurring using novel image prior constraints. Optik - International Journal for Light and Electron Optics, 124(20):4429-4434.
  28. Simoncelli, E. (1997). Statistical models for images: Compression, restoration and synthesis. In In 31st Asilomar Conf on Signals, Systems and Computers, pages 673-678. IEEE Comp. Soc.
  29. Sorel, M. and Sroubek, F. (2012). Restoration in the presence of unknown spatially varying blur. In Gunturk, B. and Li, X., editors, Image Restoration: Fundamentals and Advances, pages 63-88. CRC Press.
  30. Srivastava, A., Lee, A. B., Simoncelli, E. P., and c. Zhu, S. (2003). On advances in statistical modeling of natural images. Journal of Math. Imaging and Vision, 18:17- 33.
  31. Wang, S., Hou, T., Border, J., Qin, H., and Miller, R. (2012). High-quality image deblurring with panchromatic pixels. ACM Trans. Graph., 31(5).
  32. Wang, Z., Simoncelli, E. P., and Bovik, A. C. (2003). Multiscale structural similarity for image quality assessment. In Proc. of the 37th IEEE Asilomar Conf. on Signals, Systems and Computers.
  33. Weiss, Y. and Freeman, W. (2007). What makes a good model of natural images? In IEEE Conf. on Comp. Vision and Pattern Recognition (CVPR), pages 1-8. IEEE Comp. Soc.
  34. Whyte, O., Sivic, J., Zisserman, A., and Ponce, J. (2012). Non-uniform deblurring for shaken images. Intern. Journal of Comp. Vision, 98(2):168-186.
  35. Yuan, L., Sun, J., Quan, L., and Shum, H.-Y. (2007). Blurred/non-blurred image alignment using sparseness prior. In Proc. of the 11th Intern. Conf. on Comp. Vision (ICCV), pages 1-8.
  36. Yun-Fang, Z. (2010). Blur detection for surveillance video based on heavy-tailed distribution. In Asia Pacific Conf. on Postgrad. Research in Microelectronics and Electronics, pages 101-105.
  37. Zhong, L., Cho, S., Metaxas, D., Paris, S., and Wang, J. (2013). Handling noise in single image deblurring using directional filters. In IEEE Conf. on Comp. Vision and Pattern Recogn. (CVPR), pages 612-619. IEEE Comp. Soc.
  38. Zhou, X., Zhou, F., Bai, X., and Xue, B. (2014). A boundary condition based deconvolution framework for image deblurring. Journal of Computational and Applied Mathematics, 261:14-19.
Download


Paper Citation


in Harvard Style

Herwig J., Linder T. and Pauli J. (2014). Removing Motion Blur using Natural Image Statistics . In Proceedings of the 3rd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM, ISBN 978-989-758-018-5, pages 125-136. DOI: 10.5220/0004830201250136


in Bibtex Style

@conference{icpram14,
author={Johannes Herwig and Timm Linder and Josef Pauli},
title={Removing Motion Blur using Natural Image Statistics},
booktitle={Proceedings of the 3rd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,},
year={2014},
pages={125-136},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0004830201250136},
isbn={978-989-758-018-5},
}


in EndNote Style

TY - CONF
JO - Proceedings of the 3rd International Conference on Pattern Recognition Applications and Methods - Volume 1: ICPRAM,
TI - Removing Motion Blur using Natural Image Statistics
SN - 978-989-758-018-5
AU - Herwig J.
AU - Linder T.
AU - Pauli J.
PY - 2014
SP - 125
EP - 136
DO - 10.5220/0004830201250136