Cancer is a group of diseases involving abnormal cell growth with varying malignancy and extent across different patients. Cytological features like prominent nucleoli, nuclear enlargement, and hyperchromasia are important to the tumor pathologist in assessment of cancer malignancy from tissue biopsies. In a recent study, Yap et al.  proposed effective prominent nucleoli detectors in histopathological images and developed different feature generation methods. These methods were based on polar gradients and were used along with support vector machine (SVM) and AdaBoost for detection purposes. In this study, we benchmark the performance of these methods along with convolutional and fully connected networks for the task of distinguishing between nuclei with and without prominent nucleolus.
Proceedings of the 5th KES International Conference on Innovation in Medicine and Healthcare (KES-InMed 2017)