Liver Tumor Classification With Advanced Deep Learning Techniques

The death rate of liver patients are high due to the diagnosis of the disease in the final stage. The deep learning technique helps to classify the type of disease at an early stage. This paper uses segmentation, feature extraction, and classification of liver cancer tissues to predict the type of cancer. Here, the augmentation-based auto-encoder framework is used for DL implementation with medical dataset integration. The anticipated deep learning model is used to create a deep network of restricted graph-based Boltzmann machine (RGBM) to define the activation of hidden unit nodes in RGBM as features and emphasize quantization by grouping these features in the method of unsupervised manner. Liver tissue classification used in medical imaging is designed by a novel 3D convolution neural network and used for discriminating normal and cancerous metastatic liver tissues from weighted MRI data