PROJECT NO: FCD-2022-2575
PROJECT NAME: Cancerous Area Detection in the H&E Stained Histopathology Images of Breast Cancer
PROGRAM: Ankara University Multidisciplinary Scientific Research Project
PROJECT SUMMARY:
Thanks to the digitization of histopathology images, scientists interested in image analysis have also entered this field of study and decisions have been made based on quantitative measurements. Within the scope of the proposed project, it is aimed to detect cancerous regions on H&E stained histopathology images of breast cancer. In this way, both the workload of pathologists and the cost of data storage will be reduced. In addition, it will be a preliminary study for future studies on the segmentation of nuclei, mitotic or tubular structures in cancerous regions.
Within the scope of the proposed project, successful results were obtained in similar problems known in the literature; Mask-RCNN, YoloV5 and Transformers deep learning methods will be applied.
The original value of the project will be to create a data set with cancerous areas on the H&E stained histopathological images of breast cancer, with the work to be carried out with the Department of Pathology of Ankara University Faculty of Medicine. The data set to be created will be the first data set created on this subject throughout our country in line with our knowledge. Secondly, it has been popular in the literature recently and has achieved successful results; Mask R-CNN, YoloV5 and Transformers deep learning methods will be applied to the presented problem. These methods have achieved successful results against their competitors in studies for different purposes. Therefore, the application and evaluation of the methods specified for the proposed problem is unique since it is not found in the known literature.