Dataset on NAFLD Severity Classification with Ultrasound Liver Image & Clinical Data
Description
This dateset contains three stage of liver condition along with clinical data and demographic data on the classification of Non-Alcoholic Fatty Liver Disease (NAFLD). This dataset meant to help in medical image analysis, machine learning, and deep learning for early liver disease detection and severity assessment. The ultrasound images are acquired with standard diagnostic ultrasound equipment in B-mode. Ultrasound imaging is a widely used, and cost-effective modality for liver assessment, making it particularly suitable for large-scale screening and early detection of NAFLD. Each sample is labelled according to NAFLD status, and disease severity grades (Normal, Benign, Malignant). These labels enable both binary and multi-classification tasks. Along with images of liver, we include clinical and demographic data relevant to NAFLD diagnosis. The clinical data included with patient age, gender, body mass index (BMI), liver enzyme measurements such as alanine aminotransferase (ALT) and aspartate aminotransferase (AST), waist size, glucose, cholesterol such as (LDL,HDL, Triglycerides). This combination of data (image and Tabular) makes possible research on multi-modal learning, which improve diagnostic accuracy and model robustness. All data on our dataset are anonymized to de-identified patient's personal information and to ensure compliance with ethical research standards and data protection guidelines. The dataset is strictly provided for research and educational purposes and does not contain any information that can be used to identify individual patients. This dataset can be used for image preprocessing, classical machine learning classification, convolutional neural network (CNN)-based deep learning, disease severity grading, and comparative studies between image-only and multi-modal diagnostic approaches. the dataset Publicly accessible under the Creative Commons Attribution 4.0 International (CC BY 4.0) license.