Real-Time Deconvolution with GPU and Spark for Big Imaging Data Analysis

  • Lianyu CaoEmail author
  • Penghui Juan
  • Yinghua Zhang
Conference paper
Part of the Lecture Notes in Computer Science book series (LNCS, volume 9530)


Light sheet fluorescence microscopy (LSFM) led researchers to get optical sections of large samples, virtually without toxicity and light bleaching and with high temporal resolution, and to record the development of large, living samples with exceptionally high information content. And images observed by LSFM with high signal to noise ratio are very suited for three-dimensional reconstruction. Deconvolution reduces blurring from out-of-focus light to improve the contrast and sharpness of image, but commercial deconvolution software is slow and expensive which cannot meet the current demand. GPU is the new many-core processor with powerful floating point performance, so we parallelized the Richardson Lucy Deconvolution on the GPU. Under ensuring image quality, the implementation on the GPU runs ~30 times faster than the implementation on the CPU. For an image of size 1024 × 1024 × 25, the deconvolved time of 50 iterations on the GPU is no more than 2 s.


Deconvolution GPU Spark Manycore Imaging 


  1. 1.
    Santi, P.A.: Light Sheet fluorescence microscopy a review. J Histochem Cytochem 59(2), 129–138 (2011). doi: 10.1369/0022155410394857 CrossRefGoogle Scholar
  2. 2.
    Reynaud, E.G., Krzic, U., Greger, K., Stelzer, E.H.K.: Light sheet-based fluorescence microscopy: more dimensions, more photons, and less photodamage. HFSF J 2(5), 266–275 (2008)CrossRefGoogle Scholar
  3. 3.
    Shaw, P.J.: Comparison of wide-field/deconvolution and confocal microscopy for 3D imaging. In: Pawley, J.B. (ed.) Handbook of Biological Confocal Microscopy, pp. 373–387. Springer, Heidelberg (1995)CrossRefGoogle Scholar
  4. 4.
  5. 5.
  6. 6.
    Zaharia, M., Chowdhury, M., Franklin, M.J.: Spark: cluster computing with working sets (2010)Google Scholar
  7. 7.
    Richardson, W.H.: Bayesian-base iterative method of image restoration. JOSA 62(1), 55–59 (1972)CrossRefGoogle Scholar
  8. 8.
    Lucy, L.B.: An iterative technique for the rectification of observed distributions. Astron. J. 79, 745–754 (1974)CrossRefGoogle Scholar
  9. 9.
  10. 10.
    NVIDIA CUDA C Programming Guide.
  11. 11.
  12. 12.
    Ryoo, S., Rodrigues, C.I., Baghsorkhi, S.S.: Optimization principles and application performance evaluation of a multithreaded GPU using CUDA. In: Proceedings of the 13th ACM SIGPLAN Symposium on Principles and Practice of Parallel Programming (2008)Google Scholar
  13. 13.
  14. 14.
  15. 15.
    Griffa, A., Garin, N., Sage, D.: Hollow Bars.
  16. 16.
  17. 17.
    Ingaramo, M., York, A.G., Hoogendoorn, E., Postma, M., Shroff, H., Patterson, G.H.: Richardson-Lucy deconvolution as a general tool for combining images with complementary strengths. ChemPhySchem 15(4), 794–800 (2014)CrossRefGoogle Scholar
  18. 18.
    Liu, H., Zhang, Z., Liu, S., Liu, T., Yan, L., Zhang, T.: Richardson-Lucy blind deconvolution of spectroscopic data with wavelet regularization. Appl. Optics 54(7), 1770–1775 (2015)CrossRefGoogle Scholar
  19. 19.
    Stohl, F., Kaminski, C.F.: A joint Richardson—Lucy deconvolution algorithm for the reconstruction of multifocal structured illumination microscopy data. Methods Appl. Fluoresc. 3(1), 014002 (2015)CrossRefGoogle Scholar
  20. 20.
    Zhu, J., Chen, L., Chen, A., Luo, G., Deng, X., Liu, X.: Fast 3D dosimetric verifications based on an electronic portal imaging device using a GPU calculation engine (2015)Google Scholar
  21. 21.
    Sherry, M., Shearer, A.: IMPAIR: massively parallel deconvolution on the GPU. In: Proceedings of SPIE - The International Society for Optical Engineering (2013)Google Scholar

Copyright information

© Springer International Publishing Switzerland 2015

Authors and Affiliations

  1. 1.School of Computer and Communication EngineeringUniversity of Science and Technology BeijingBeijingChina
  2. 2.Institute of Computing TechnologyChinese Academy of SciencesBeijingChina

Personalised recommendations