Critical Points Detection in Neuron Microscopy Images
-
1
Erasmus MC, Medical Informatics and Radiology, Netherlands
Measuring the morphology of neuronal cells is an essential step towards understanding neuronal cell and network functionality. Fluorescence microscopy is a powerful tool for capturing detailed information about neuronal cell morphology and connectivity. The enormous amount of image data acquired in typical experiments is not used to its full potential due to the fact that automated neuron reconstruction methods are still very far from being perfect and expert manual image annotation is too laborious. Hence the development of computational image analysis methods that allow accurate and efficient neuron reconstruction is imperative [1-3]. Neurons are tree-like structures whose accurate representation depends critically on the bifurcations and end-points. Automatic identification of these critical points in the images provides important clues for neuronal reconstruction.
We have developed a novel method that automatically detects and characterizes bifurcations [4] as well as end-points in microscopy images of neuronal cells. Several challenges emerge when processing such images, such as nonuniform intensity (caused by inhomogeneous staining), the complexity and diversity of the image structures, and the fact that many of these structures are below the optical resolution limit. To address these issues, our method combines a newly developed directional filtering algorithm and a fuzzy-logic [5] rule-based reasoning scheme to decide about the presence and the type of critical point at each image location. We have carefully designed a set of fuzzy-logic rules that leverages the local image context at each point to make accurate decisions.
The developed method has been successfully applied for critical points detection in preliminary experiments involving fluorescence microscopy image data sets from various labs in order to test robustness. The presented results (see attached figure) illustrate the potential of the method (with bifurcations and end-points shown as red and yellow dots, respectively, where larger diameters indicate higher reliability). We are in the process of developing a fully automated system for neuronal reconstruction that takes advantage of the sparseness of the image data and at the same time exploits local image context to achieve both high efficiency and high accuracy. The presented critical points detection method provides essential input to the system and will prove useful to existing neuron tracing methods as well.
References
[1] E. Meijering. Neuron Tracing in Perspective. Cytometry Part A 77(7):693-704, July 2010.
[2] D. E. Donohue and G. A. Ascoli. Automated Reconstruction of Neuronal Morphology: An Overview. Brain Research Reviews 67(1-2):94-102, June 2011.
[3] Y. Liu. The DIADEM and Beyond. Neuroinformatics 9(2-3):99-102, September 2011.
[4] M. Radojevic, I. Smal, W. Niessen, E. Meijering. Fuzzy Logic Based Detection of Neuron Bifurcations in Microscopy Images. Proceedings of the IEEE International Symposium on Biomedical Imaging, May 2014, in press.
[5] J. M. Mendel. Fuzzy Logic Systems for Engineering: A Tutorial. Proceedings of the IEEE 83(3):345-377, March 1995.
Keywords:
neuron reconstruction,
Fuzzy Logic,
Flourescence microscopy,
Image Analysis and Processing,
Neurobiology
Conference:
Neuroinformatics 2014, Leiden, Netherlands, 25 Aug - 27 Aug, 2014.
Presentation Type:
Poster, to be considered for oral presentation
Topic:
Computational neuroscience
Citation:
Radojevic
M,
Smal
I,
Niessen
W and
Meijering
E
(2014). Critical Points Detection in Neuron Microscopy Images.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2014.
doi: 10.3389/conf.fninf.2014.18.00057
Copyright:
The abstracts in this collection have not been subject to any Frontiers peer review or checks, and are not endorsed by Frontiers.
They are made available through the Frontiers publishing platform as a service to conference organizers and presenters.
The copyright in the individual abstracts is owned by the author of each abstract or his/her employer unless otherwise stated.
Each abstract, as well as the collection of abstracts, are published under a Creative Commons CC-BY 4.0 (attribution) licence (https://creativecommons.org/licenses/by/4.0/) and may thus be reproduced, translated, adapted and be the subject of derivative works provided the authors and Frontiers are attributed.
For Frontiers’ terms and conditions please see https://www.frontiersin.org/legal/terms-and-conditions.
Received:
24 Apr 2014;
Published Online:
04 Jun 2014.
*
Correspondence:
Mr. Miroslav Radojevic, Erasmus MC, Medical Informatics and Radiology, Rotterdam, Netherlands, miroslav.radojevic@gmail.com