Building an Al model to improve the efficiency of parasite inspection services for foreigners employed

Kaohsiung Municipal Siaogang Hospital

2024 National Healthcare Quality Award

  Highlights of Workload Reduction  

1. Manual microscopic examination requires 5.4 minutes per sample. With the integration of an Al-driven model, image scanning and analysis can be completed within just 1 minute, significantly enhancing efficiency and enabling rapid result acquisition.

2. For a cohort of 500 individuals, the report issuance time was reduced from 55 minutes prior to implementation to just 3.8 minutes, demonstrating a substantial improvement in operational efficiency.

 

This system utilizes a modified convolutional neural network (CNN) architecture, based on Faster R-CNN, to identify protozoan parasite eggs from image scans. With AI model interpretation, each image can be analyzed within one minute, providing rapid identification of the parasite egg type.

Source: 2025 Application of Digital Healthecare

Back to Collections