Classification and Segmentation of Cells in Anatomic & Time Lapse Microscopic Images based on Geometrical Features and Machine Learning
-
1
kuLeuven, Belgium
-
2
IMEC, Belgium
Classification and segmentation of astrocytes in Anatomic & Time lapse microscopic images has become increasingly important during last few years. In order to quantify cells in an objective manner, we followed a twofold approach: on one hand to address image noise, we extract wide range of geometrical features using Laplacian of Gaussian operator (Duits 2003), Gaussian Derivative operator, Bi-Laplacian of Gaussian operator and scale representation generated by Anisotropic decomposition of the Laplacian (Anisotropic filter, ALoG Filter (Weickert 1998)& (Lindberg 1994)). On the other hand to exploit the geometrical features we propose a semi-supervised active learning technique that will allow us to interactively collect new training examples in a robust manner involving little human attention. The initial learning model is built with few training examples and used to predict the unseen training instances. The user will provide feedback on this prediction. This process is continued till user is satisfied with the outcomes. After that, the model will be used to predict the unseen testing images. We are implementing an interactive framework in ImageJ using Weka as learning framework. The resulting framework is general enough to learn models for different kind of tissues and dendritic trees etc and able to adapt new learning task. The technique is successfully applied to two-photon images and comparisons are done with threshold based learning approaches, where the learning model is sensitive to global or a local threshold parameters. The extracted features are tested with wide range of machine learning classifiers like Support Vector machines, Naive Bayes Classifier etc to check the generality of features.
References
Duits R, M. Felsberg, L. Florack, and B. Platel. (2003). Alpha-scale spaces on a bounded domain. Scale Space Methods in Computer Vision, 494- 510. Springer.
Lindeberg, T. (1994). Scale-space theory: A basic tool for analysing structures at different scales. J App Stat. 21 (2): 224–270.
Weickert J. (1998). Anisotropic Diffusion in Image Processing. ECMI Series. Teubner-Verlag, Stuttgart.
Keywords:
scale-space,
machine learning,
Active Learning,
Cell segmentation,
two-photon imaging
Conference:
Neuroinformatics 2016, Reading, United Kingdom, 3 Sep - 4 Sep, 2016.
Presentation Type:
Poster
Topic:
Computational neuroscience
Citation:
Vohra
SK and
Prodanov
D
(2016). Classification and Segmentation of Cells in Anatomic & Time Lapse Microscopic Images based on Geometrical Features and Machine Learning.
Front. Neuroinform.
Conference Abstract:
Neuroinformatics 2016.
doi: 10.3389/conf.fninf.2016.20.00050
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:
30 Apr 2016;
Published Online:
18 Jul 2016.
*
Correspondence:
Mr. Sumit K Vohra, kuLeuven, leuven, Belgium, sumit.3203@gmail.com
Dr. Dimiter Prodanov, IMEC, Leuven, Belgium, dimiterpp@gmail.com