Non-Invasive Diabetes Detection Using Facial Texture Features Captured in a Less Restrictive Environment

The prevalence of Diabetes Mellitus (DM) worldwide has risen dramatically with 1 of 3 deaths in Western Pacific (WP) region according to the 2017 report of International Diabetes Federation (IDF). The Philippines ranks fifth in WP with the most cases of diabetes. Local experts and IDF estimate that...

Full description

Saved in:
Bibliographic Details
Main Author: Garcia, Christina
Format: text
Published: Archīum Ateneo 2019
Online Access:https://archium.ateneo.edu/theses-dissertations/417
Tags: Add Tag
No Tags, Be the first to tag this record!
Institution: Ateneo De Manila University
Description
Summary:The prevalence of Diabetes Mellitus (DM) worldwide has risen dramatically with 1 of 3 deaths in Western Pacific (WP) region according to the 2017 report of International Diabetes Federation (IDF). The Philippines ranks fifth in WP with the most cases of diabetes. Local experts and IDF estimate that half of the people with diabetes are unaware they have it and will likely remain undiagnosed. Though non-invasive DM detection techniques have gained consideration of more researchers, presently they have restrictive set-up for image capture. This paper explores the performance of using mobile device as convenient tool for image capture of DM and healthy dataset for non-invasive detection using facial block texture features and Gabor filter. Filipino participants that undergo regular check-ups for diabetes monitoring were chosen within the age inclusion criteria of 20 to 69 years old in which surveys for Philippines assessed the occurrence of diabetes to be most prevalent. For each subject, a mobile device 12mp and 7mp cameras, and laptop camera were used to take the photo approximately 30cm in front of the face under normal lighting condition to ensure full coverage and avoid unnecessary background. A ratio of 70:30 training to testing set was maintained and extracted facial blocks were classified using SVM and KNN. A total of 100 images from each camera were captured, pre-processed, filtered and iterated to compare performance of data. Performance of the system was measured in terms of Accuracy, Specificity and Sensitivity. 96 cases from 3 cameras, 4 methods of texture feature extraction, 2 classifiers, and 4 iterations of dataset were compared. Best performance of 96.7% accuracy, 100% sensitivity and 93% specificity were achieved from 12mp back camera using SVM with 100 images.