Non-rigid Registration for Large Sets of Microscopic Images on Graphics Processors
- 227 Downloads
Microscopic imaging is an important tool for characterizing tissue morphology and pathology. 3D reconstruction and visualization of large sample tissue structure requires registration of large sets of high-resolution images. However, the scale of this problem presents a challenge for automatic registration methods. In this paper we present a novel method for efficient automatic registration using graphics processing units (GPUs) and parallel programming. Comparing a C++ CPU implementation with Compute Unified Device Architecture (CUDA) libraries and pthreads running on GPU we achieve a speed-up factor of up to 4.11× with a single GPU and 6.68× with a GPU pair. We present execution times for a benchmark composed of two sets of large-scale images: mouse placenta (16K ×16K pixels) and breast cancer tumors (23K ×62K pixels). It takes more than 12 hours for the genetic case in C++ to register a typical sample composed of 500 consecutive slides, which was reduced to less than 2 hours using two GPUs, in addition to a very promising scalability for extending those gains easily on a large number of GPUs in a distributed system.
KeywordsMicroscopic imaging Image registration and segmentation Pattern analysis Feature detection Graphics processors High-performance computing
This work was partially supported by the Ministry of Education of Spain (TIC2003-06623, PR-2007-0014), Junta de Andalucía of Spain (P06-TIC-02109), US NIH grant R01 DC06458-01A1 and the startup fund from the Department of Biomedical Informatics at the Ohio State University, US.
We thank Dr. Gustavo Leone from the Ohio State University Cancer Center for providing us the images from mouse placenta and mouse mammary gland we used during the experiments outlined in this paper. We also thank Dr. Dennis Sessanna and Dr. Donald Stredney from the Ohio Supercomputing Center for providing us access to the BALE visualization cluster where most of our execution times were obtained.
- 4.Huijismans, D., Lamers, W., Los, J., & Strackee, J. (1986). Toward computerized morphometric facilities: A review of 58 software packages for computer-aided three-dimensional reconstruction, quantification, and picture generation from parallel serial sections. The Anatomical Record, 216(4), 449–470.CrossRefGoogle Scholar
- 5.Moss, V. (1989). The computation of 3-dimensional morphology from serial sections. European Journal of Cell Biology, 48, 61–64.Google Scholar
- 7.Hajnal, J., Derek, H., & Hawkes, D. (2001). Medical image registration. Boca Raton: CRC.Google Scholar
- 8.Goshtasby, A. (2005). 2-D and 3-D image registration: For medical, remote sensing, and industrial applications. New York: Wiley-Interscience.Google Scholar
- 12.Hill, W., & Baldock, R. (2003). The constrained distance transform: Interactive atlas registration with large deformations through constrained distance. In Proceedings of the workshop on image registration in deformable environments.Google Scholar
- 13.Yoo, T. (2004). Insight into images: Principles and practice for segmentation, registration, and image analysis. AK, Peters.Google Scholar
- 17.Cooper, L., Huang, K., Sharma, A., Mosaliganti, K., & Pan, T. (2006). Registration vs. reconstruction: building 3-d models from 2-d microscopy images. In Proceedings of the workshop on multiscale biological imaging, data mining and informatics (pp. 57–58).Google Scholar
- 18.Huang, K., Cooper, L., Sharma, A., & Pan, T. (2006). Fast automatic registration algorithm for large microscopy images. In Proceedings of the IEEENLM life science systems & applications workshop (pp. 1–2).Google Scholar
- 19.Koshevoy, P., Tasdizen, T., & Whitaker, R. (2006). Implementation of an automatic slice-to-slice registration tool. University of Utah, SCI Institute Technical Report UUSCI-2006-018. (Online) Available: http://www.sci.utah.edu/publications/SCITechReports/UUSCI-2006-018.pdf.
- 20.Prescott, J., Clary, M., Wiet, G., Pan, T., & Huang, K. (2006). Automatic registration of large set of microscopic images using high-level. In Proceedings of the IEEE international symposium on medical imaging (pp. 1284–1287).Google Scholar
- 21.Mosaliganti, R., Pan, T., Sharp, R., Ridgway, R., Iyengar, S., Gulacy, A., et al. (2006). Registration and 3d visualization of large microscopy images. In Proceedings of the SPIE medical imaging meeting (pp. 6144:923–934).Google Scholar
- 23.Botnen, M., & Ueland, H. (2004). The GPU as a computational resource in medical image processing. Dept. of Computer and Information Science, Norwegian Univ. of Science and Technology, Tech. Rep.Google Scholar
- 24.Owens, J. D., Luebke, D., Govindaraju, N., Harris, M., Kruger, J., Lefohn, A. E., et al. (2007). A survey of general-purpose computation on graphics hardware. Journal of Computer Graphics Forum, 26, 21–51.Google Scholar
- 26.Lewis, J. P. (1995). Fast normalized cross-correlation. In Vision interface. Canadian image processing and pattern recognition society (pp. 120–123). (Online) Available: citeseer.ist.psu.edu/lewis95fast.html.
- 27.Compute Unified Device Architecture (CUDA) (2007). Home page maintained by Nvidia. http://developer.nvidia.com/object/cuda.html. Accessed 1 May 2008.
- 28.GPGPU (2007). A web site dedicated to the general-purpose on the GPU. http://www.gpgpu.org.
- 29.Fatica, M., Luebke, D., Buck, I., Owens, D., Harris, M., Stone, J., et al. (2007). Cuda tutorial at supercomputing 2007 (November). Accessed 28 Dec 2007.Google Scholar
- 30.CUFFT library (2007). Home page maintained by nvidia. http://developer.download.nvidia.com/compute/cuda/1_1/CUFFT_Library_1.1.pdf. Accessed 28 Dec 2007.
- 31.The FFTW library (2007). FFTW home page. http://www.fftw.org. Accessed 1 May 2008.
- 32.Cooper, L., Naidu, S., Leone, G., Saltz, J., & Huang, K. (2007). Registering high resolution microscopic images with different histochemical stainings - a tool for mapping gene expression with cellular structures. In Proceedings of the workshop on microscopic image analysis with applications in biomedicine.Google Scholar
- 37.Ohara, M., Yeo, H., Savino, F., Iyengar, G., Gong, L., Inoue, H., et al. (2007). Real time mutual information-based linear registration on the cell broadband engine processor. In Proceedings of the IEEE international symposium on medical imaging (ISBI ) (pp. 33–36).Google Scholar
- 38.Fan, Z., Qiu, F., Kaufman, A., & Yoakum-Stover, S. (2006). GPU cluster for high performance computing. In Proceedings 2004 ACM/IEEE intl. conference for high performance computing, networking, storage and analysis (pp. 47–53). Washington DC, USA.Google Scholar
- 41.Ino, F., Gomita, J., Kawasaki, Y., & Hagihara, K. (2006). A GPGPU approach for accelerating 2-d/3-d rigid registration of medical images. In Proceedings of the 4th international symposium on parallel and distributed processing and applications (ISPA) (pp. 769–780). Lecture Notes in Computer Science 4331. Berlin: Springer.Google Scholar
- 43.Guha, S., Krisnan, S., & Venkatasubramanian, S. (2005). Data visualization and mining using the GPU. In tutorial at 11th ACM international conference on knowledge discovery and data mining (KDD 2005).Google Scholar
- 44.Hadwiger, M., Langer, C., Scharsach, H., & Buhler, K. (2004). State of the art report on GPU-based segmentation. VRVis Research Center, Vienna, Austria, Tech. Rep. TR-VRVIS-2004-17.Google Scholar
- 45.Fatahalian, K., Sugerman, J., & Hanrahan, P. (2004). Understanding the efficiency of GPU algorithms for matrix-matrix multiplication. In Proceedings of the ACM SIGGRAPH - EUROGRAPHICS workshop on graphics hardware (HWWS’04). Grenoble, France (August).Google Scholar
- 46.Moreland, K., & Angel, E. (2004). The FFTW on a GPU. In Proceedings of the ACM SIGGRAPH - EUROGRAPHICS workshop on graphics hardware (HWWS’03). San Diego, California, USA (August).Google Scholar
- 47.TESLA (2008). GPGPU high-end hardware solutions from Nvidia. http://www.nvidia.com/object/tesla_computing_solutions.html. Accessed 1 Jan 2008.
- 48.FireStream (2008). GPU hardware solutions from AMD/ATI. http://ati.amd.com/products/streamprocessor/specs.html. Accessed 1 Jan 2008.
- 49.The BALE Supercomputer at the Ohio Supercomputer Center (OSC). http://www.osc.edu/supercomputing/hardware.