Breast fibroadenoma automatic detection using k-means based hybrid segmentation method

Fibroadenoma is a benign tumor that has some features similar to a malignant one. The aim of this study was to examine the impact of fibroadenoma cases on the results of the automatic breast cancer diagnostic system based on the quantitative morphometric analysis of fine needle biopsy microscopic images. The database of 50 patients (500 images) of benign and malignant lesions used previously in our research was enriched by an additional 25 patients (250 images) of fibroadenoma cases. Experiments were performed using the k-means based hybrid segmentation method. The system was tested on a set of real case medical images with promising results.