Nucleus Segmentation in Breast Histopathology Images
Balakrishnan Lakshmanan, Snekalatha Saravanakumar · 2018 International Conference on Current Trends towards Converging Technologies (ICCTCT) · 2018
Today Cancer growth evaluating is a vital research region in Image Processing. As of late the picture handling systems are utilized generally in a few therapeutic zones for enhancing prior recognition and recovering stages, it needs the time which is essential to find the malady in the patient as conceivable as quick, particularly in different cancer tumors, like breast cancer. Automated nucleus detection is troublesome in Mitotic nuclei come out as hyper chromatic objects without a clear nuclear membrane in H and E (Hematoxyline and Eosin) in histopathology images. Extracting poor contrast features and falsely detect tumour cells present in breast images from single modality image is the problem. The shape of nucleus is varies for different phases. The identification procedure moves toward becoming time-consuming and to a great degree troublesome because of huge assortment of shapes, size and low recurrence of nuclei undergoing mitosis is the major problem that leads to false positive results in detection process of nuclei. We proposed a system for segmenting both affected and non affected cells in Histopathology images using Otsu's thresholding and Gabor features extraction.