Home > Events > Oral Candidacy - Xiao (Sean) Bo

Oral Candidacy - Xiao (Sean) Bo

Start: 5/17/2018 at 9:00AM
End: 5/17/2018 at 12:00PM
Location: 258 Fitzpatrick
Attendees: Faculty and students are welcome to attend the presentation portion of the defense.
Add to calendar:
iCal vCal

Xiao (Sean) Bo

Oral Candidacy

May 17, 2018        9:00 am         258 Fitzpatrick

Adviser:  Dr. Christian Poellabauer

Committee Members:

Dr. Alan Huebner      Dr. Gregory Madey      Dr. Dong Wang


Reliability of Sensor Data for Human Action Recognition


Action recognition is frequently used to detect an individual's motion patterns based on a series of observations of the individual's body and environment. As more sensors (e.g., GPS and accelerometers) are integrated into modern smartphones, interest in action recognition using mobile devices has increased in recent years, where machine learning plays a central role in the action detection. The accuracy of supervised machine learning algorithms for action recognition depends on the data quality, which can be impaired by incorrect user-provided activity labels. The purpose of this research is to provide novel techniques to measure the impact of label errors on data quality and to detect and remove erroneous labels. To measure data quality, this work will apply and adapt the concept of intraclass correlation coefficient to measure the consistency of time-series data (such as accelerometer readings). Further, the work will study the use of deep neural networks and transfer learning for label error detection.