Fully automatic classification of breast cancer microarray images

Abstract

A microarray image is used as an accurate method for diagnosis of cancerous diseases. The aim of this research is to provide an approach for detection of breast cancer type. First, raw data is extracted from microarray images. Determining the exact location of each gene is carried out using image processing techniques. Then, by the sum of the pixels associated with each gene, the amount of “genes expression” is extracted as raw data. To identify more effective genes, information gain method on the set of raw data is used. Finally, the type of cancer can be recognized via analyzing the obtained data using a decision tree. The proposed approach has an accuracy of 95.23% in diagnosing the breast cancer types.

Keywords

• Microarray
• Gene expression
• Information gain
• Breast cancer
• Decision tree