The Classification Platform Applied to Mammographic Images
In this paper is tackled the mammographic image classification problem. From the previous developed CLAP – CLAssification Platform for use with Matlab, several computational paradigms emphasizing neural networks, support vector machines and fuzzy systems,
- PDF / 146,343 Bytes
- 10 Pages / 439.37 x 666.142 pts Page_size
- 55 Downloads / 209 Views
The Classification Platform Applied to Mammographic Images P.J.S. Gonc¸alves
Abstract In this paper is tackled the mammographic image classification problem. From the previous developed CLAP – CLAssification Platform for use with Matlab, several computational paradigms emphasizing neural networks, support vector machines and fuzzy systems, were used to classify mammographic images in two classes, i.e., with or without tumour. To perform the classification task, features must be extracted from the mammographic images. Amongst the methods implemented in CLAP, features obtained from the co-occurrence matrix and wavelets were used, to describe the texture of the region of interest in the image. Results obtained while training and validating the mentioned computational paradigms, show that support vector machines outperform the other two types of classifiers, independently of the features selected.
22.1
Introduction
The CLAP – CLAssification Platform [1], is a software tool designed to classify images. CLAP can extract features from images in order to organize the images, e.g., mammographic images with or without tumour. CLAP is suitable to obtain data for classification purposes using images of any type, and then classify them. This paper applies the CLAP to breast cancer diagnosis to help medical doctors in diagnostic tasks, using mammographic images. In this paper, mammograms were classified in two classes: with tumour and without tumour.
P.J.S. Gonc¸alves (*) School of Technology, Polytechnic Institute of Castelo Branco, Av. Empresa´rio, Castelo Branco 6000-767, Portugal IDMEC/LAETA, Technical University of Lisbon (TU Lisbon), Av. Rovisco Pais, Lisbon 1049-001, Portugal e-mail: [email protected] A. Madureira et al. (eds.), Computational Intelligence and Decision Making: Trends and 239 Applications, Intelligent Systems, Control and Automation: Science and Engineering 61, DOI 10.1007/978-94-007-4722-7_22, # Springer Science+Business Media Dordrecht 2013
240
P.J.S. Gonc¸alves
From the diagnostic data, mammograms, the software tool can propose to the medical doctor a possible assessment. The basis of the tool is twofold. The first part is related to image processing algorithms that outputs the descriptors used in the second part, which is related to classification. The first part of the CLAP extracts the image features from the captured image, a mammogram. The algorithms used in this paper are the wavelet decomposition for texture and the gray-level co-occurrence matrix (GLCM). These algorithms give a large number of features that are used in the second part of the system. The features to be used can be chosen by the user. The second part of the system uses the image features obtained in the image processing part of the system. The soft computing systems implemented in the platform are based in fuzzy logic, neural networks and support vector machines. From fuzzy logic theory, the system model was built using takagi-sugeno fuzzy models, where clustering is performed using Fuzzy C-Means and Gustafson-Kessel
Data Loading...
