Convolutional neural network application for cells segmentation in immunocytochemical study

HER2/neu status of breast cancer is important for treatment strategy choice, but nowadays it is evaluated manually by pathologist. The automation of this procedure is an urgent task, because it will allow to free a pathologist from routine work. But the problem of cells segmentation is difficult for classical methods of computer vision due to the frequent presence of erythrocytic background and non-cellular elements at specimen' image. In this work we propose segmentation algorithm, based on deep convolutional neural networks. Is it has been shown that using this approach, it's possible to get better results, than using classical computer vision algorithms.