James Patrick A. Acang | Doreen D. Domingo | Enoch Caryl M. Taclan | Donna Mae B. Fronda
Chronic binge alcohol consumption is a leading cause of chronic liver disease worldwide. This disease could lead to cirrhosis and hepatocellular carcinoma. Prolonged alcohol drinking significantly affects the liver, kidney, pancreas, heart, lungs, Central Nervous System (CNS), and other organs, as reported by recent work in literature. Building a tool for assessing and quantifying these damages could provide a different perspective on medi-cal diagnosis. This work is part of the effort of infusing computational biology in medical diagnosis. In this research, we investigated how to detect the Central Vein and Sinusoids automatically by a computer. Based on the literature, the Central Vein and Sinusoids are a few of the liver morpho-anatomical parts affected by alcohol. Detecting these parts automatically could revolutionize the damage estimation in the liver. We composed our dataset from the Michigan Histology and Virtual Microscopy Learning Resource. One hundred histopathological images were collected and were processed for analysis. These histopathological images were manually cropped using 40x.svs (Scan Scope Virtual Slide) magnification and were manually labeled and annotated for ground truth comparison and performance analysis. Sixty percent (60%) of the dataset was used for training, and the other forty percent (40%) were used for testing. Gaussian and thresholding filters were used for the recognizer. Our initial results show that the model could meet 90%-pixel accuracy in detecting the Central Vein and Sinusoids in the Histopathological images. This finding shows promising development in the field of medical diagnosis.
ISSN 2719-1990 (Print)
Philippine E-Journals
