Emory University Uses Oracle Cloud To Help Fight Parkinson's Disease

A 2022 Parkinson's Foundation study revealed that nearly 90,000 Americans are diagnosed with Parkinson's disease (PD) annually, with the total number expected to reach 1.2 million by 2030. Traditional methods for assessing PD symptoms are outdated and subjective.

Researchers at Emory University aimed to develop a machine-learning system to measure PD symptoms more accurately using videos and wearable devices. They faced challenges with computational resources but found a solution with Oracle Cloud Infrastructure (OCI), enabling faster compute speeds and HIPAA compliance. OCI's capabilities allowed the team to run multiple GPUs in parallel, significantly accelerating their research. Their neural network achieved high precision in matching expert ratings for PD symptoms.

Learn how the project will benefit the Human Activity Recognition research community by providing a large, labeled motion capture and video dataset, encouraging further ML advancements, and releasing pretrained ML models to aid in advanced PD analysis by accessing the full case study below.