A METHOD FOR QUANTITATIVE DESCRIPTION OF BIOMEDICAL IMAGES BASED ON SUPERPIXEL DICTIONARIES
Abstract
With this study, a method for quantitative description of biomedical images based on splitting the target image into superpixels followed by categorization using precalculated superpixel dictionaries is proposed. The method has been tested on the tasks of recognition of biomedical images of three types: lung CT images, histology images of ovary and thyroid tissues. The results of the experiments performed suggest that the method proposed may provide recognition performance comparable or better than when using conventional methods of texture description.
About the Authors
V. A. Liauchuk
Объединенный институт проблем информатики НАН Беларуси
Belarus
Belarus
V. A. Kovalev
Объединенный институт проблем информатики НАН Беларуси
Belarus
Belarus
References
1. Kovalev, V. Color Co-occurence Descriptors for Querying-by-Example / V. Kovalev, S. Volmer // Proc. of the 1998 Conf. on MultiMedia Modeling. – Switzerland, 1998. – P. 32–38.
2. Content-based image retrieval as a method for melanoma diagnosis / V. Liauchuk [et al.] // Proc. of the 29th Intern. Congress on Computer Assisted Radiology and Surgery (CARS– 2015). – Spain, 2015. – Vol. 10. – P. 291–292.
3. Pinto, N. A high-throughput screening approach to discovering good forms of biologically inspired visual representation / N. Pinto, D. Doukhan, D. Cox // PLoS Computational Biology. – 2009. – Vol. 5(11). – P. 1–12.
4. Evaluating Bag-of-Visual-Words Representations in Scene Classification / J. Yang [et al.] // Proc. of the Intern. Workshop on Multimedia Information Retrieval. – Germany, Augsburg, 2007. – P. 197–206.
5. Quantized Image Patches Co-occurrence Matrix: A New Statistical Approach for Texture Classification using Image Patch Exemplars / Zh. Liu [et al.] // Proc. of the 3th Intern. Conf. on Digital Image Processing (ICDIP 2011). – China, Chengdu, 2011. – Vol. 8009. – P. 56–64.
6. Fulkerson, B. Class segmentation and object localization with superpixel neighborhoods / B. Fulkerson, A. Vedaldi, S. Soatto // Proc. of the 12th Intern. Conf. on Computer Vision. – Kyoto, Japan. – 2009. – P. 8.
7. Levinshtein, A. Multiscale symmetric part detection and grouping / A. Levinshtein, C. Sminchisescu, S. Dickinson // Proc. of the 12th Intern. Conf. on Computer Vision. – Kyoto, Sapan. – 2009. – Vol. 104. – P. 117–134.
8. Li, X. An Object Co-occurrence Assisted Hierarchical Model for Scene Understanding / X. Li, Y. Guo // Proc. of the British Machine Vision Conference. – Surrey, 2012. – P. 1–11.
9. Multi-scale feature learning on pixels and super-pixels for seminal vesicles MRI segmentation / Q. Gao [et al.] // SPIE 9034, Medical Imaging 2014: Image Proc. – 2014. – Vol. 9034. – Р. 903407–903407-6.
10. Micusik, B. Semantic segmentation of street scenes by superpixel co-occurrence and 3D geometry / B. Micusik, J. Kosecka // IEEE Workshop on Video-Oriented Object and Event Classification (VOEC). – Japan, 2009. – P. 625–632.
11. He, X. Learning and incorporating top-down cues in image segmentation / X. He, R. Zemel, D. Ray // Proc. of the 9th European Conf. on Computer Vision. – Austria, 2006. – P. 338–351.
12. Sicre, R. SuperPixel based angular differences as a mid-level image descriptor / R. Sicre, E. Tasli, T. Gevers // Proc. of the 22th Intern. Conf. on Pattern Recognition (ICPR). – Sweden, 2014. – P. 3732–3737.
13. Content-based image retrieval of histology images / A. Dmitruk [et al.] // Proc. of Swedish Symposium on Image Analysis (SSBA-2010). – Uppsala, Sweden, 2010. – P. 91–92.
14. Tighe, J. SuperParsing: Scalable nonparametric image parsing with superpixels / J. Tighe, S. Lazebnik // ECCV’10 Proc. of the 11th European Conf. on Computer Vision. – Germany, Heidelberg, 2010. – P. 352–365.
15. SLIC Superpixels Compared to State-of-the-art Superpixel Methods / R. Achanta [et al.] // IEEE Transactions on Pattern Analysis and Machine Intelligence. – 2012. – Vol. 34(11). – P. 2274– 2282.
16. Ковалев, В.А. Влияние мер близости в пространстве признаков на качество поиска медицинских изображений по содержанию / В.А. Ковалев, А.А. Дмитрук // Информатика. – 2011. – № 30. – С. 5–11.
17. Efron, B. Bootstrap Methods: Another Look at the Jackknife / B. Efron // Annals of Statistics. – 1979. – Vol. 1. – P. 1–26.
18. Matas, T. Extended set of local binary patterns for rapid object detection / T. Matas, J. Matas // Proc. of the Computer Vision Winter Workshop. – Czech Republic, Nové Hrady, 2010. – Vol. 2010. – P. 37–43.
Review
For citations:
Liauchuk V.A., Kovalev V.A. A METHOD FOR QUANTITATIVE DESCRIPTION OF BIOMEDICAL IMAGES BASED ON SUPERPIXEL DICTIONARIES. Informatics. 2016;(1):49-57. (In Russ.)
Views: 750
Email this article 