Research Article
Dehydration Scan: An Artificial Intelligence Assisted Smartphone-Based System for Early Detection of Dehydration
@INPROCEEDINGS{10.1007/978-3-031-34586-9_20, author={Priyeta Saha and Syed Muhammad Ibne Zulfiker and Tanzima Hashem and Khandker Aftarul Islam}, title={Dehydration Scan: An Artificial Intelligence Assisted Smartphone-Based System for Early Detection of Dehydration}, proceedings={Pervasive Computing Technologies for Healthcare. 16th EAI International Conference, PervasiveHealth 2022, Thessaloniki, Greece, December 12-14, 2022, Proceedings}, proceedings_a={PERVASIVEHEALTH}, year={2023}, month={6}, keywords={Dehydration detection Mobile image analysis Deep learning}, doi={10.1007/978-3-031-34586-9_20} }- Priyeta Saha
Syed Muhammad Ibne Zulfiker
Tanzima Hashem
Khandker Aftarul Islam
Year: 2023
Dehydration Scan: An Artificial Intelligence Assisted Smartphone-Based System for Early Detection of Dehydration
PERVASIVEHEALTH
Springer
DOI: 10.1007/978-3-031-34586-9_20
Abstract
Dehydration occurs due to fluid loss from the human body, affects regular body functions, and causes health complications. Physical exercises, poor fluid intake, and diseases like fever and diarrhea may result in dehydration. Current clinical and laboratory-based dehydration detection techniques are expensive, time-consuming, and require people to visit medical facilities, which often do not exist in destitute areas. Though recent research has focused on monitoring physiological parameters (e.g., heart rate, stress, and oxygen) and detecting diseases using smartphones, the area of dehydration detection has not been sufficiently addressed. We present a smartphone-based early dehydration detection system using artificial intelligence, which is ubiquitous, quick, and does not require any additional cost or expertise to operate. We develop a siamese network-based deep learning model to detect the changes in the facial landmarks that appear from dehydration and are not detectable with the naked eyes of general people. Our model provides an overall accuracy of 76.1% and is lightweight enough to run on a smartphone processor. By integrating it in the background, we develop a smartphone app, “Dehydration Scan” that simply captures facial images of individuals and detects their hydration status. Knowing early about dehydration allows people to take oral rehydration solutions and avoid severe dehydration.
