Biomedical Image Segmentation by Deep Learning Methods

Abstract

Deep learning methods have been employed to predict and analyse various application in medical imaging. Deep Learning technology is a computational algorithm that learns by itself to demonstrate a desired behaviours. Neural network processes the input neurons according to the corresponding types of networks based on algorithm provided and passes it to the hidden layer. Finally, it outputs the result through output layer. Deep learning algorithms tend to be more useful in different applications. It plays important role in biomedical image segmentations such as identifying skin cancer, lung cancer, brain tumour, skin psoriasis, etc. Deep learning includes algorithms like Convolutional Neural Network (CNN), Restricted Boltzmann Machine (RBM), Generative Adversarial Network (GAN), Recurrent Neural Network (RNN), U-Net, V-net, Fully Convolutional Attention Network (FCANET), Docker- powered based deep learning, ResNet18, ResNet50, SqueezeNet and DenseNet-121 which processes on medical images and helps in identifying the defect in earlier stage by helping the physician to start the treatment process. This paper is about the review of deep learning algorithms using medical image segmentation. Future implementations can be performed through additional feature for the existing algorithm with better performance.