Deep learning model-based segmentation of medical diseases from MRI and CT images

Medical image segmentation is quite challenging field. Deep Learning (DL) based Unet model is used for medical image segmentation. The Unet architecture is based on encoder and decoder which is the most successful method. Unet based methods still have a drawback that is not able to fully utilize the output features of the node's convolutional units. This paper presented deep learning model-base segmentation of medical diseases from MRI and CT scan images data with the help of 2D-Unet and 3D-Unet model. The model using package nibabel for reading, visualizing (itk, itkweidgets, ipywidgts), and 3D-Unet method for classification and segmentation of MRI and CT scan images. The proposed system has been tested on Medical Segmentation Decathion (MSD) datasets for the data of Brain, Spleen and Heart. The performance metrices has been analysis by F1-score (F1), Intersection over Union (IOU), and Dice Factor coefficient (DFC) on Brain, Spleen and Heart datasets. The proposed model is outperformed as comparison with some recent network model.