Inspiration

We currently don't have a reliable non-invasive method to screen, identify and track diabetes patients. Voice analysis is promising non-invasive method for tracking diabetes because hardware is readily available and voice parameters have shown promising association with diabetes condition. Recent research has indicated the promise of using voice parameters to detect diseases. In particular, recent research has indicated that vocal characteristics can be used to distinguish people with type 2 diabetes from healthy individuals. A study with over 177 voice samples showed that voice parameters (i.e. jitter, shimmer, perturbation quotient, noise to harmonic ratio) have significant difference in their values for type 2 diabetes patients as compared to non-diabetic persons (DOI: 10.1109/AEEICB.2016.7538402)

What it does

We propose an application that uses microphones and voice activation technology built into standard phones and consumer devices (i.e. devices like Aamzon Alexa) that allows diagnosis and tracking or type 2 diabetes using voice characteristics. This type of solution could be a ground-breaking technology for non-invasive tracking of diabetes as well as helping people with the condition monitor their disease.

How I built it

Multidimensional voice program analysis (MDVP) is used for acoustic analysis of voice samples and categorize and track diabetes condition. Such technology can be embedded into devices and provide a method for ambient screening and identification of people with diabetes in various scenarios like doctors office and Telehealth calls. This technology could also be useful for at hoe tracking of diabetes condition.

Team

Our multidisciplinary team has 4 years experience in pharmaceutical chemistry and 9 years experience in development of signal processing algorithms.