Abstract
In this paper, a new medical image classification scheme is proposed using self-organizing map (SOM) combined with multiscale technique. It addresses the problem of the handling of edge pixels in the traditional multiscale SOM classifiers. First, to solve the difficulty in manual selection of edge pixels, a multiscale edge detection algorithm based on wavelet transform is proposed. Edge pixels detected are then selected into the training set as a new class and a multiscale SOM classifier is trained using this training set. In this new scheme, the SOM classifier can perform both the classification on the entire image and the edge detection simultaneously. On the other hand, the misclassification of the traditional multiscale SOM classifier in regions near edges is greatly reduced and the correct classification is improved at the same time.
Similar content being viewed by others
References
Xuan J, Adali T, Wang Y. Segmentation of magnetic resonance brain image: Integrating region growing and edge detection. InProc. ICIP’95, Washington D.C., 1995, pp. 544–547.
Wells W M, Grimson W E L, Kikinis R, Jolesz F A. Adaptive, segmentation of MRI Data.IEEE Trans. Medical Imaging, Aug. 1996, 15(4): 429–442.
Kamber M, Collins D L, Francis G S, Evans A C. Model-based 3-D segmentation of multiple sclerosis lesions in magnetic resonance brain images.IEEE Trans. Medical Imaging, Sept. 1995, 14(3): 442–453.
Kass M, Witkin A, Terzopoulos D. Snakes: Active contour models.Int. J. Comput. Vision, 1987, 1: 321–331.
Ozkan M, Dawant B M, Maciunas R J. Neural-network-based segmentation of multi-modal medical images: A comparative and prospective study.IEEE Trans. Medical Imaging, Sept. 1993, 12(3): 534–544.
Kohonen T. The self-organizing map. InProceedings of the IEEE, 1990, 78(9): 1464–1480.
Schenone A, Firenze F, Acquarone Fet al. Segmentation of multivariate medical images via unsupervised clustering with ‘Adaptive Resolution’.Computerized Medical Imaging and Graphics, 1996, 20(3): 119–129.
Busch C, Grob M H. Interactive Neural Network Texture Analysis and Visualization for Surface Reconstruction in Medical Imaging. InEUROGRAPHICS’93, University of York, UK, 1993, pp. 49–60.
Florack L M J, ter Haar Romeny B M, Koenderink J J, Viergever M A. Scale and the differential structure of images.Image and Vision Computing, July/Aug., 1992, 10(6): 376–388.
Haring S, Viergever M A, Kok J N. A multiscale approach to image segmentation using Kohonen networks. InIPMI, Proc. the 13th Conference, Gmitro A, Barrett H H (eds.), Springer-Verlag, Berlin, 1993, pp. 212–224.
Drebin Robert A, Loren Carpenter, Pat Hanrahan. Volume rendering.Computer Graphics, Aug. 1988, 22(4): 65–74.
Mallat S G, Zhong, S. Characterization of signals from multi-scale edges.IEEE Trans. Pattern Analysis and Machine Intell., 1992, 14: 710–723.
Mallat Set al. Wave 2. ftp://ftp.cs.nyu.edu/pub/software/wave/wave2.tar.Z
Author information
Authors and Affiliations
Additional information
This project is supported by the National Natural Science Foundation (No. 69775001) and the National Hi-Tech Program (863-05-18) of China.
Ma Feng is a Ph.D. candidate in the Department of Automation, Tsinghua University, P.R. China. He received his B.S. and M.S. degrees of engineering from the Department of Automation, Tsinghua University in 1993 and 1996 respectively. His research interests are image processing especially medical image segmentation, pattern recognition, neural network, multiscale techniques including wavelet theory and scale-space theory.
Xia Shaowei is a Professor in the Department of Automation, Tsinghua University. Her research areas are system engineering, neural network and pattern recognition.
Rights and permissions
About this article
Cite this article
Ma, F., Xia, S. A multiscale approach to automatic medical image segmentation using self-organizing map. J. of Comput. Sci. & Technol. 13, 402–409 (1998). https://doi.org/10.1007/BF02948498
Received:
Revised:
Issue Date:
DOI: https://doi.org/10.1007/BF02948498