Automated Counting of Cells in Breast Cytology Images Using Level Set Method

Statistical analysis of cells in breast cytology images is very important for the diagnosis of various diseases in the female population in developed and developing countries. Manual detection and counting of the cancer cell in real time is not only difficult but hugely time-consuming for pathologists. In this paper, we propose an algorithm for automatic analysis of breast cytology using Fine Needle Aspiration Cytology (FNAC) images. The proposed technique uses statistical measures which include perceptual information (like color) and morphological characteristics for the estimation of the initial cell boundary. Similarly, the level set technique is used for efficient and accurate identification of cellular objects which help in the precise counting of individual cancer cell breast cytology images. Experimental results obtained during the demonstration of the proposed approach show high correlations in precision with manual counting by a pathologist. It has been proved that the proposed approach is efficient in processing cells for counting the cancerous cells with high accuracy and avoid discrepancies(like color variations, human error) in manual counting by a pathologist.