Performance Evaluation of Automated Algorithm for Breast Cancer Cell Counting
Chalit Primkhajeepong, Pornchai Phukpattaranont, Somchai Limsiroratana, Pleumjit Boonyaphiphat, Kanita Kayasut · International Journal of Computer and Electrical Engineering · 2010
This paper presents an automated algorithm for breast cancer cell counting and its performance.The algorithm for analyzing stained breast cancer cell image consists of four procedures, i.e. image preprocessing, segmentation, feature extraction, and classification.In the image preprocessing, the wavelet transform is performed.The global thresholding and morphological operations are performed in segmentation.In the feature extraction, the average of b* in CIE L*a*b* color space is extracted.The segmented cells are classified by using extracted feature.If the average of b* is positive, the cancer cell is positive cell.In addition, if the average of b* is negative, the cancer cell is negative cell.The segmentation results show that the average performance is 85% when square-shaped structuring element and city block distance transform are used.The classification results show that the average performance is 94%.