Abstract
Connected Component Labeling (CCL) is a well-known algorithm with many applications in image processing and computer vision. Given the growth in terms of inter-pixel relationships and the amount of information stored in a single pixel, the time to run CCL analysis on an image continues to increase rapidly. In this paper we present an accelerated version of CCL using NVIDIA’s Compute Unified Device Architecture (CUDA) framework to address this growing overhead. Our parallelization approach decomposes CCL while respecting all global dependencies across the image. We compare our implementation against serial execution and parallelized implementations developed on OpenMP. We show that our parallelized CCL algorithm targeting NVIDIA’s CUDA can significantly increase performance, while still ensuring labeling quality.
This is a preview of subscription content, log in via an institution to check access.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
Preview
Unable to display preview. Download preview PDF.
References
Zhao, H.L., Fan, Y.B., Zhang, T.X., Sang, H.S.: Stripe-based connected components labelling. Electronics Letters 46(21), 1434–1436 (2010)
Oliveira, V., Lotufo, R.A.: A study on connected components labeling algorithms using GPUs. XXIII Sibgrapi, Graphics, Patterns and Images (2010)
Bailey, D., Johnston, C.: Singles Pass Connected Components Analysis. Image and Vision Computing (2007)
Klaiber, M., Rockstroh, L.Z., Wang, B.Y., Simon, S.: A memory-efficient parallel single pass architecture for connected component labeling of streamed images. Field-Programmable Technology (FPT), 159–165, 10–12 (2012)
Paralic, M.: Fast connected component labeling in binary images. In: 35th Telecommunications and Signal Processing (TSP), vol. 709, pp. 3–4 (2012)
Hwu, W.-M.: GPU Computing Gems Emerald Edition. M. Kaufmann (2011)
NVIDIA’s Next Generation CUDA Compute Architecture Whitepaper: Kepler GK110. Nvidia (2013)
Foley, J.: Migrating your code from Tesla Fermi to Tesla K20X, with examples from QUDA Lattice QDC library. Microway, Inc. (2013)
National Electrical Manufacturers Association: Digital Imaging and Communications in Medicine (DICOM)., http://medical.nema.org/standard.html
Mehta, S., Misra, A., Singhal, A., Kumar, P., Mittal, A., Palaniappan, K.: Parallel implementation of video surveillance algorithms on GPU architectures using CUDA. In: 17th IEEE Int. Conf. Advanced Computing and Communications, ADCOM (2009)
Riha, L., Manohar, M.: GPU accelerated one-pass algorithm for computing minimal rectangles of connected components, pp. 479–484. IEEE Computer Society Press (2011)
Author information
Authors and Affiliations
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2014 Springer International Publishing Switzerland
About this paper
Cite this paper
Nina Paravecino, F., Kaeli, D. (2014). Accelerated Connected Component Labeling Using CUDA Framework. In: Chmielewski, L.J., Kozera, R., Shin, BS., Wojciechowski, K. (eds) Computer Vision and Graphics. ICCVG 2014. Lecture Notes in Computer Science, vol 8671. Springer, Cham. https://doi.org/10.1007/978-3-319-11331-9_60
Download citation
DOI: https://doi.org/10.1007/978-3-319-11331-9_60
Publisher Name: Springer, Cham
Print ISBN: 978-3-319-11330-2
Online ISBN: 978-3-319-11331-9
eBook Packages: Computer ScienceComputer Science (R0)