Work

EPFH-Net: efficient pyramid forward harmonic network for non-alcoholic fatty liver disease detection

Year: 2026
Type: article
Abstract: Detecting Non-Alcoholic Fatty Liver Disease (NAFLD) involves identifying fat accumulation in the liver but not linked to excessive alcohol consumption. NAFLD is a common liver disorder where fat builds up in liver cells, potentially leading to inflammation, liver damage, or liver cancer. Liver biopsy is the therapy used to detect NAFLD in the olden days, which consumed more time and provided inaccurate results. Therefore, the new framework named Efficient Pyramid Forward Harmonic Network (EPFH-Net) is introduced to detect NAFLD effectively. First, the ultrasound liver image sourced from the B-mode fatty liver ultrasound images dataset is fed as the input. Next, the extraction of the region is done by the Region of Interest (ROI) and the features, such as Convolutional Neural Network (CNN), Gray Level Co-occurrence Matrix (GLCM), Pyramid Histogram of Oriented Gradients (PHOG), and Center Symmetric-Local Binary Patterns (CS-LBP), are extracted. Then, Non-alcoholic Fatty Liver Disease (NAFLD) is detected using EPFH-Net, which is a combination of Efficient Network (EfficientNet), Pyramid Network (PyramidNet), and harmonic analysis. Finally, the detected outputs are fatty or normal. The experimental outcomes attained by EPFH-Net are 92.12% of accuracy, 92.52% of TPR, and 91.88% of TNR for the k-value of 9.
Language: English
Subfield: Hepatology
Field: Medicine
Open Access status: gold
Estimated APC paid: $2,690
OpenAlex ID: W7234047212
FWCI:
Cites: 26
Cited by:
Related to: 10