Smart Eye Drops Sponsored by Vodafone Marco Aguilar

Smart Eye Drops Sponsored by Vodafone Marco Aguilar, Aaron Alvarez, Ana Cortes, Jonathan Nunez, Ivan Yu, Feng You Advisor: Navid Amini, Kevin Delao Liaison: Haley Kirk

Project Goals To build an eye drop adherence application that will allow physicians to: • Monitor patient activity • View analytics with an intuitive and user-friendly interface • Adjust regimen or prescription dosage based on the data we collect • Increase eye drop adherence rates Feng You

Project Significance • Physicians will be able to make personalized prescription regimens for their patients • Will enable physicians to analyze whether a treatment plan is working or not • Allowing doctors more control outside their office helps treatment for diseases like glaucoma • Cut back on medical costs Feng You

Project Requirements Marco Aguilar

Project Requirements • Develop a completely new eye drop adherence monitor • Accurately record when patients take medication • Provide valuable information to physicians to help them make better decisions regarding their patients Marco Aguilar

DFD 0 Marco Aguilar

DFD 1 Marco Aguilar

Technologies Ana Cortes

Technologies Used • MBient. Lab API and Technologies • Amazon Web Services (AWS) • Android OS for mobile application • Weka (Machine Learning tool) Ana Cortes

MMR 1" round force sensor MMC 1. 5" square force sensor Technologies Used • MBient. Lab Meta. Motion. C (MMC) and Meta. Motion. R (MMR) offer real-time information of motion and environmental sensor data • Pressure sensor and MMC are attached to the eye drop container to collect input from the user and state (e. g. tilted or at equilibrium) Ana Cortes

Mobile Development Aaron Alvarez

Meta. Wear • Meta. Wear Sensor • Multiple small sensors such as • Accelerometer (calculate acceleration) • Gyro (calculate rotation) • Others • Pins (install external sensors) • Force sensor (calculate pressure) • Meta. Wear API • Tools that allows us to communicate with the sensors • Gather data from sensors Aaron Alvarez

Mobile Application • Android studio • Mobile Application • Allows us to use Meta. Wear API • External APIs to help use Visualize Data Cleanly • Bluetooth connectivity • Meta. Wear MMC • Smart eye drop that is constantly monitoring Aaron Alvarez

Mobile Application • Meta. Wear sensor (generate data) • Meta. Wear API (gather data) • Android Studio (visualize data) Aaron Alvarez

Machine Learning Jonathan Nunez

Weka (Machine Learning Tool) • Collection of Machine Learning algorithms • Provides tools for data classification, clustering, regression, and visualization • Provides data analysis and predictive modeling for our eye drop adherence application Jonathan Nunez

Machine Learning Classification • Naive Bayes, support vector machine, random forest, and deep learning • Machine learning algorithm: may get required results with less data • Deep learning: more accurate results as more data is gathers Jonathan Nunez

Cloud Service Ivan Yu

Eye Drop Adherence Cycle Data sent to Machine Learning Algorithms Real-Time Data Patient's regimen or dosage is adjusted (If needed) Analytics sent to Physician's device Ivan Yu

• Why not use the in-house cloud service (Metacloud)? • Meta. Cloud does not support machine learning Why AWS? • Microsoft Azure? Google Cloud Services (GCS)? • Amazon made the most sense for our project • Amazon is a lot easier to scale into than GCS • While pricier than Azure, it has an easier interface to use • Sagemaker • Has a lot of documentation and examples of Machine Learning Ivan Yu

Successfully connect different sensors Accomplishments Retrieve data from sensors and display in real time Ivan Yu

Collect live data from sensors Goals for next semester Have a fully functioning prototype Connect to AWS to display data Implement Machine Learning Ivan Yu

Demo