AzureVisual - 2019

Quantifying Tremors and Progression of Parkinson’s disease through an Android App and Bluetooth Low Energy Technology​
Abstract

Currently, the diagnosis of Parkinson’s Disease is being hindered by the lack of proper biomarkers that exist in other diseases. Clinical symptoms often are validated when there is over 60% loss of dopamine in the body. This is simply too late for a treatment process to begin. As a result, the development of an effective detection and monitoring system is a top priority. In this project, a completely novel tool was created using a Convolutional Pose Estimation machine (CPM), that allows for accurate and effective detection and monitoring of Parkinson’s Disease. Using advanced feature extraction algorithms, proper attributes were inputted to a CPM, allowing for a classification/diagnosis to be outputted.  After proper training and evaluation, the CPM was used in a highly advanced Android App that can be used by doctors, patients, and caregivers. The diagnostic model reached an accuracy greater than 95%, higher than any currently released literature. The developed diagnostic tool eliminates the requirement of expensive infrastructure, and instead utilizes the discovered biomarkers to provide a novel, objective approach to PD diagnosis that will aid in the development of pathogenesis-targeted therapeutics, as well as progressive monitoring of Parkinson’s Disease, allowing for effective data collection and analysis over time. 

Project Snapshot
Awards
  • Founders’ Award for Excellence in Public Health for the project – AzureVisual – Creating a method using Artificial Intelligence to detect Parkinson’s disease in the early stages using Visual Diagnostics (2019)
  • First Place at the regional science fair
Photos and Videos