Akademik Veri Yönetim Sistemi
29th IEEE Conference on Signal Processing and Communications Applications (SIU), ELECTR NETWORK, 9 - 11 Haziran 2021
The number of breast cancer cases and mortality rate has been on the rise globally. For effective treatment and curbing the mortality rate, early diagnosis is paramount. To evaluate the effectiveness of treatment, a fundamental step is histopathological examination of the state of nuclei. However this is a challenging task for various reasons. Accordingly, we perform nuclei detection in breast cancer histopathology images using RetinaNet with a ResNet-152 as the feature extractor. The data used for training and validation of the model was carefully collected at our partner teaching hospital. A mean Average Precision (mAP) of 78% was achieved by optimizing the Focal Loss. Thus outperforming current state-of-the-art methods which have been reported in the literature.