Applying GPU parallel programming for image processing and clustering

Cover Page

Cite item

Full Text

Open Access Open Access
Restricted Access Access granted
Restricted Access Subscription or Fee Access

Abstract

This paper presents state-of-the-art image processing and structural analysis software tools that use GPU parallel programming to achieve substantial performance gains. The software suite combines advanced preprocessing techniques, object identification methods, clustering algorithms, and analysis tools to facilitate efficient and precise analysis of complex imaging datasets. The case studies illustrate the software’s versatility and effectiveness across diverse scientific domains, including materials science, biological research, and astronomy. By exploiting GPU parallel programming, the tools deliver performance improvements of 5–20x compared to traditional sequential programming, enabling real-time visualization and expedited data processing. The intuitive user interface empowers researchers to fine-tune parameters, visualize results, and interpret data with ease, streamlining the research workflow. The broader impacts of these tools include accelerating scientific discovery, enhancing data analysis accuracy, and driving innovation across diverse scientific fields. A notable example of their effectiveness is the processing and analysis of electron microscopy images of amorphous alloys. The developed algorithms and software tools demonstrate promising results in this area, facilitating detailed studies of atomic structure and degree of orderliness.

Full Text

Restricted Access

About the authors

Dagim Sileshi Dilla

Far Eastern Federal University

Author for correspondence.
Email: dilla.d@dvfu.ru
ORCID iD: 0000-0002-9100-1257
SPIN-code: 7200-1921

postgraduate student, Institute of Mathematics and Computer Technologies; research engineer, Electron Microscopy and Imaging Laboratory

Russian Federation, Vladivostok

Evgeniy V. Pustovalov

Far Eastern Federal University

Email: pustovalov.ev@dvfu.ru
ORCID iD: 0000-0003-1036-3975
SPIN-code: 6192-2432

Dr. Sci. (Phys.-Math.); Professor, Department of Information and Computer Systems, Institute of Mathematics and Computer Technologies; Head of the educational program 09.03.02 “Information systems and technologies”, profile “Programming of robotic systems”

Russian Federation, Vladivostok

Irina L. Artemeva

Far Eastern Federal University

Email: artemeva.il@dvfu.ru
ORCID iD: 0000-0003-2088-5259
SPIN-code: 8161-1313

Dr. Sci. (Eng.), Professor; deputy Director for Scientific Works, Institute of Mathematics and Computer Technology

Russian Federation, Vladivostok

References

  1. Dyson M.A. Advances in computational methods for transmission electron microscopy simulation and image processing. Abstract of dis. University of Warwick. URL: http://go.warwick.ac.uk/wrap/729437
  2. Pryor A., Ophus C., Miao J. A streaming multi-GPU implementation of image simulation algorithms for scanning transmission electron microscopy. Advanced Structural and Chemical Imaging. 2017. No. 3 (15). doi: 10.1186/s40679-017–0048-z.
  3. Roy S., Prabhat L.Q., Tran L., Ang L.K. Parallel programming for electron microscopy image processing // Proceedings of the International Conference on Parallel Processing. ACM, 2019. Pp. 1–9.
  4. Xu Y., Li W., Fu Y. et al. Parallel implementation of RELION on GPU-accelerated clouds for efficient single-particle cryo-EM. Journal of Structural Biology. 2021. No. 213 (3). P. 107762.
  5. Jian L., Wang C., Liu Y. et al. Parallel data mining techniques on Graphics Processing Unit with Compute Unified Device Architecture (CUDA). J. Supercomput. 2013. No. 64. Pp. 942–967. doi: 10.1007/s11227-011-0672-7.
  6. Gault B., Moody M.P., Cairney J.M., Ringer S.P. Atom probe microscopy. Springer Science & Business Media, 2012. doi: 10.1007/978-1-4614-3436-8.
  7. Kirkland E.J. Advanced computing in electron microscopy. Springer, 2010. 289 p.
  8. Ronneberger O., Fischer P., Brox T. U-net: Convolutional networks for biomedical image segmentation. In: International Conference on Medical image computing and computer-assisted intervention. Springer Cham, 2015. Pp. 234–241. doi: 10.1109/TPDS.2020.2975562.
  9. Scheres S., Nunez-Ramirez R., Sorzano C. et al. Image processing for electron microscopy single-particle analysis using XMIPP. Nat. Protoc. 2008. No. 3. Pp. 977–990. doi: 10.1038/nprot.2008.62.
  10. Pustovalov E.V., Plotnikov V.S., Grudin B.N. et al. Electron tomography algorithms in scanning transmission electron microscopy. Bull. Russ. Acad. Sci. Phys. 2013. No. 77. Pp. 995–998. doi: 10.3103/S1062873813080340
  11. Sorzano C.O.S., Vilas J.L., Cuenca-Alba J. et al. Statistical image analysis in electron microscopy. In: Biophysics and the challenges of life. Springer, Cham. 2018. Pp. 79–98.
  12. Dilla D.S., Pustovalov E.V., Fedorets A.N. Advanced electron microscopy image processing for analyzing amorphous alloys: Electron Microscopy Image Cluster Analyzer (EMICA). Tool and results. Computational Nanotechnology. 2024. Vol. 11. No. 1. Pp. 104–111. doi: 10.33693/2313-223X-2024-11-1-104–111. EDN: DYNPTQ.
  13. Dilla D.S., Pustovalov E.V., Fedorets A.N., Frolov A.M. Exploring amorphous alloys: Advanced electron microscopy and cluster analysis. Computational Nanotechnology. 2024. Vol. 11. No. 1. Pp. 112–120. doi: 10.33693/2313-223X-2024-11-1-112–120. EDN: DYUGCI.
  14. Zhu Y., Ouyang Q., Xu Y. A deep convolutional neural network approach to single-particle recognition in cryo-electron microscopy. BMC Bioinformatics. 2017. No. 18 (1). Pp. 1–9. doi: 10.1186/s12859-017–1757-y.
  15. Ophus C. A fast image simulation algorithm for scanning transmission electron microscopy. Adv. Struct. Chem. Imag. 2017. No. 3. P. 13. doi: 10.1186/s40679-017-0046-1.

Supplementary files

Supplementary Files
Action
1. JATS XML
Download
2. Fig. 1. Modular data analysis architecture: main architectural modules illustrating the data flow from raw data to visualization

Download (323KB)
Indexing metadata
3. Fig. 2. Scatterplot diagram of the NiW sample, showing the identified atomic particle clusters (HRTEM)]

Download (250KB)
Indexing metadata
4. Fig. 3. Linearized Kullback–Leibler (K-L) divergence for the CDF (Div (SP(B/V)) and Lebesgue measure (Div (Mu(B/V)) of the bond-to-vertex ratio, as a function of the ordering level in atomic structure images

Download (91KB)
Indexing metadata
5. Fig. 4. Bond and vertex ratio analysis, showing clusters with a higher degree of ordering as their size increases

Download (107KB)
Indexing metadata
6. Fig. 5. Radial distribution function (RDF) for amorphous alloys, showing particle distance probabilities and revealing local near-order

Download (127KB)
Indexing metadata
7. Fig. 6. Atomic density distribution (a) and radial distribution function (b) for two simulated realizations of the FeB25 alloy

Download (149KB)
Indexing metadata
8. Fig. 7. Evolution of the angle distributions between bonds for different degrees of ordering, from 10 to 90%, in clusters consisting of more than 3 vertices

Download (206KB)
Indexing metadata
9. Fig. 8. Change in the proportion of clusters with certain angles between bonds

Download (211KB)
Indexing metadata