Huang, Yi, and Song, Insu (2022) PhysioVec: IoT Biosignal Based Search Engine for Gastrointestinal Health. In: Proceedings of the 7th International Conference on Computational Intelligence and Applications. pp. 230-236. From: ICCIA 2022: IEEE 7th International Conference on Computational Intelligence and Applications, 24-26 June 2022, Nanjing, China.

PDF (Published Version) - Published Version Restricted to Repository staff only

Abstract

Gastrointestinal problems are major health threats to term newborn babies. There are currently no known methods for monitoring the gastrointestinal health of these babies in ICU units contributing to thousands of yearly mortality rates in Australia alone. The internet and Health Social networks (HSN) provide a large amount of useful information for patients. However, finding the right information on HSN is time-consuming and challenging because data from HSN is too large to be processed manually. We develop PhysioVec, a Bowel-Sound IoT to HSN search engine that extracts physiological measurements from bowel sounds providing an automated search of HSN. PhysioVec consists of three parts: Local Recurrent Transformer (LRT), a Multivariate radial-basis Logistic Interpreter (MLI), and a sentence embedding module. LRT combines local attention and recurrent Transformer encoder to reduce overfitting and improve the performance of bowel sound segmentation. The physiological measurements extracted from bowel sounds are used to search for relevant health information on the internet. PhysioVec achieved 100.00% precision in the top one search results for bowel sound with both vomiting and bowel obstruction. Our proposed framework allows patients and doctors to search for useful information in HSN by continuously monitoring bowel sounds with minimal discomfort.

Actions (Repository Staff Only)

Item Control Page