Date

2019 12:00 AM

Major

Computer Science and Mathematics

Department

Computer Science and Mathematics

College

Liberal Arts and Sciences

Project Advisor

Eric William Davis

Description

A common feature of smart watch devices is the ability to measure heart rate. This is done via a process called photoplethysmography, where green light is shone through the skin, and absorbed by the red colored blood cells. The change in how much light is reflected back by the blood allows the sensors in the watch to measure its wearer's heart rate. This method is prone to errors. If the watch is moved slightly, it could cause the sensors to not receive the light that is reflected back. These errors are most commonly reflected in drastic spikes or dips in heart rate. The goal of this research is to classify low integrity data points present in heart rate data. We also attempt to repair erroneous heart rate data in order. Heart rate data can be used to calculate various metrics, such as Total Daily Energy Expenditure (TDEE) and Basal Metabolic Rate (BMR). These metrics are useful for tracking fitness, so being able to measure them using data from a personal device is incredible valuable. Thus, these metrics are used in this research to show how repairing the data can have a great improvement on measurements.

File Format

application/pdf

Share

COinS
 
Jan 1st, 12:00 AM

Classification and Repair of Low Integrity Heart Data

A common feature of smart watch devices is the ability to measure heart rate. This is done via a process called photoplethysmography, where green light is shone through the skin, and absorbed by the red colored blood cells. The change in how much light is reflected back by the blood allows the sensors in the watch to measure its wearer's heart rate. This method is prone to errors. If the watch is moved slightly, it could cause the sensors to not receive the light that is reflected back. These errors are most commonly reflected in drastic spikes or dips in heart rate. The goal of this research is to classify low integrity data points present in heart rate data. We also attempt to repair erroneous heart rate data in order. Heart rate data can be used to calculate various metrics, such as Total Daily Energy Expenditure (TDEE) and Basal Metabolic Rate (BMR). These metrics are useful for tracking fitness, so being able to measure them using data from a personal device is incredible valuable. Thus, these metrics are used in this research to show how repairing the data can have a great improvement on measurements.