A Machine Learning Approach Towards Fatty Liver Disease Detection In Liver Ultrasound Images
Adarsh Kuzhipathalil, Anto Thomas, Keerthana Chand, Elmer Jeto Gomes Ataide, Alexander Link, Annika Niemann, Sylvia Saalfeld, Michael Friebe, Jens Ziegle
Otto von Guericke Universität Magdeburg, Faculty of Computer Science
Abstract
Fatty liver disease (FLD) is one of the prominent diseases which affects the normal functionality of the liver by building vacuoles of fat in the liver cells. FLD is an indicator of imbalance in the metabolic system and could cause cardiovascular diseases, liver inflammation, cirrhosis and furthermore neoplasm. Detection and specification of a FLD are beneficial to arrange an early and best adapted treatment. We present a computer aided diagnostic (CAD) tool for FLD detection using ultrasound images. Therefore we developed a model pipeline with separate segmentation and classification modules. During the development phase these modules were trained on 6 patient cases and validated with 2. The whole model was evaluated on a totally different set of data with 5 patient cases and performed with an overall classification accuracy of 0.84. The model showed impressive performance considering the size of training data. Also the multi-module architecture enables the predictions from the model to be better explainable.