A Combined Neural Network and Mathematical Morphology Method for Computerized Detection of Microcalcifications

In this paper we present a new method for the automatic detection of microcalcifications combining morphological operations and artificial neural networks (ANN). The input chosen is a whole digitalized mammogram while the output of the algorithm is a new mammogram where microcalcifications (clustered or not) appear to be highlighted in white colour individually. Two new filters have been designed: one is based on mathematical morphology combined with other segmentation techniques and statistical methods, and the other one is made by training an ANN in order to be able to differentiate between those pixels belonging to a microcalcification and those being normal parenchymal patterns. Results for both algorithms are separately shown and how the combination of these two approaches improves the results.

[1]  Minoru Okada.  Extraction of User Entered Components from A Personal Bankcheck Using Morphological Subtraction , 1997, Int. J. Pattern Recognit. Artif. Intell..

[2]  Robert M. Pap,et al.  Handbook of neural computing applications , 1990 .

[3]  Gilles Privat,et al.  Asynchronous Relaxation of Morphological Operators: A Joint Algorithm-Architecture Perspective , 1997, Int. J. Pattern Recognit. Artif. Intell..

[4]  Huai Li,et al.  Artificial convolution neural network for medical image pattern recognition , 1995, Neural Networks.

[5]  K Doi,et al.  Computerized detection of clustered microcalcifications in digital mammograms: applications of artificial neural networks. , 1992, Medical physics.

[6]  Daniel Peña Sánchez de Rivera.  Estadística: modelos y métodos , 2000 .

[7]  C. Metz.  ROC Methodology in Radiologic Imaging , 1986, Investigative radiology.

[8]  Kunio Doi,et al.  Neural network approach for differential diagnosis of interstitial lung diseases , 1990, Medical Imaging: Image Processing.

[9]  J M Boone,et al.  Neural networks in radiologic diagnosis. I. Introduction and illustration. , 1990, Investigative radiology.

[10]  G. P. Cohen,et al.  Characteristics of breast carcinomas missed by screening radiologists. , 1997, Radiology.

[11]  N Roberts,et al.  Segmentation and numerical analysis of microcalcifications on mammograms using mathematical morphology. , 1997, The British journal of radiology.

[12]  J. Crespo.  Morphological connected filters and intra-region smoothing for image segmentation , 1993 .