Diagnosis of Breast Cancer Using Deep Regulatory Learning (UNET)

Different patients, all with the same type of cancer, have different reactions to one type of treatment, sodoctors are at a crossroads and do not know how to determine the treatment that works best. In this study,first, data related to patients have been collected. A model has also been developed to categorize imagesof these patients. In this model, Unet network based on deep supervisory learning is used. In thisnetwork, a prediction is made for each pixel and the original input location information remains. Finally,pixel-by-pixel categorization is performed on the layer output feature map to achieve pixel-levelcategorization. This model can accept input images of any size because there is no fully connected layer.In addition, because the model does not use image blocks around the pixel as input during training andforecasting, segmentation accuracy is improved and storage overhead is reduced.

Keywords:

Deep learning algorithms , convulsive neural networks , breast cancer , layer Nervous