Fatty Liver Disease Stage Classification Using Ultra Sound Images by ResNet Deep Learning Model

Abstract

Fatty liver disease (FLD) is a growing global health concern, with its prevalence on the rise due to factors such as sedentary lifestyles and poor dietary choices. The non-invasive classification of FLD stages using ultrasound images is a vital diagnostic endeavor that aids in early detection and intervention, thus mitigating the potential for severe liver-related complications. In this research, a ResNet-based deep learning framework is proposed for ultrasound images based automatic staging of FLD. The model uses features obtained through ultrasound scans to classify the stage of FLD, i.e Grade 1, Grade 2 and Grade 3. They are trained, validated and tested on a dataset of standardized ultrasound images. This indicates the model's ability to classify FLD at a stage correctly and proposes a potential avenue for improved efficacy in diagnostics in hepatology.