Association Technique based on Classification for Classifying Microcalcification and Mass in Mammogram

Currently, mammography is recognized as the most effective imaging modality for breast cancer screening. The challenge of using mammography is how to locate the area, which is indeed a solitary geographic abnormality. In mammography screening it is important to define the risk for women who have radiologically negative findings and for those who might develop malignancy later in life. Microcalcification and mass segmentation are used frequently as the first step in mammography screening. The main objective of this paper is to apply association technique based on classification algorithm to classify microcalcification and mass in mammogram. The system that we propose consists of: (i) a preprocessing phase to enhance the quality of the image and followed by segmentating region of interest; (ii) a phase for mining a transactional table; and (iii) a phase for organizing the resulted association rules in a classification model. This paper also illustrates how important the data cleaning phase in building the data mining process for image classification. The proposed method was evaluated using the mammogram data from Mammographic Image Analysis Society (MIAS). The MIAS data consist of 207 images of normal breast, 64 benign, and 51 malignant. 85 mammograms of MIAS data have mass, and 25 mammograms have microcalcification. The features of mean and Gray Level Co-occurrence Matrix homogeneity have been proved to be potential for discriminating microcalcification from mass. The accuracy obtained by this method is 83%.

[1]  Saroj Kumar Lenka,et al.  A REVIEW ON COMPUTER AIDED MAMMOGRAPHY FOR BREA ST CANCER DIAGNOSIS AND CLASSIFICATION USING IMAGE MINING METHODOL OG Y , 2011 .

[2]  Jiawei Han,et al.  CPAR: Classification based on Predictive Association Rules , 2003, SDM.

[3]  Musa Mohd Mokji,et al.  Adaptive Thresholding Based on Co-occurrence Matrix Edge Information , 2007, First Asia International Conference on Modelling & Simulation (AMS'07).

[4]  Dr. med Marton Lanyi.  Diagnosis and Differential Diagnosis of Breast Calcifications , 1987, Springer Berlin Heidelberg.

[5]  Han Zhang,et al.  Automated breast masses segmentation in digitized mammograms , 2004, IEEE International Workshop on Biomedical Circuits and Systems, 2004..

[6]  Gunilla Svane.  Screening mammography : breast cancer diagnosis in asymptomatic women , 1993 .

[7]  F. Parmiggiani,et al.  An investigation of the textural characteristics associated with gray level cooccurrence matrix statistical parameters , 1995, IEEE Transactions on Geoscience and Remote Sensing.

[8]  Osmar R. Zaïane,et al.  Mammography Classification By an Association Rule-based Classifier , 2002, MDM/KDD.

[9]  R. Besar,et al.  Biomedical Imaging and Intervention Journal Identification of Masses in Digital Mammogram Using Gray Level Co-occurrence Matrices , 2022 .

[10]  John B. Shoven,et al.  I , Edinburgh Medical and Surgical Journal.

[11]  M Le Gal,et al.  [Diagnostic value of clustered microcalcifications discovered by mammography (apropos of 227 cases with histological verification and without a palpable breast tumor)]. , 1984, Bulletin du cancer.

[12]  Robert J. Pizzutiello,et al.  Introduction to Medical Radiographic Imaging , 1993 .

[13]  Mody Univesity.  Image Mining for Mammogram Classification by Association Rule Using Statistical and GLCM features , 2011 .

[14]  Spyros G. Skiadopoulos,et al.  A Texture Analysis Approach for Characterizing Microcalcifications on Mammograms , 2006 .