Liver Tumor Segmentation and Classification Using Deep Learning Methods

One of the most common causes of death due to cancer is liver cancer. After several types of research, automatically segmenting and classifying liver tumors remains a challenging task due to the deformable shape in CT, MERI, and ultrasound images and the low tissue contrast between surrounding organs. A liver tumor must be detected early to plan a treatment strategy, confirm the diagnosis with accuracy, and acquire a thorough knowledge of the tumor to determine its seriousness. In recent years, many researchers identified various methods for accurate prediction by utilizing segmentation and classification approaches. This research represents the methodologies like 3Dimensitional Deep Convolutional Neural Network (3D DCNN), Adding Inception Module-Unet (AIM-Unet), Simple Linear Iterative Clustering-based Deep Graph Network (SLIC-DGN), OPTimized Guided Contrast Enhancement (OPTGCE), Gradational modular network (GraMNet), Hyper-parameter tuned Improved Deep Neural Network (HI-DNN), Hybrid ResUNet, Unified learning multi-task model network (ULM-net) are utilized for liver tumor disease segmentation and classification analysis. This survey investigates the benefits, drawbacks, and relations between segmentation and classification-based prediction techniques.