Computer Science Location: Link Lab 211
Add to Calendar 2022-12-12T15:00:00 2022-12-12T15:00:00 America/New_York Ph.D. Proposal Presentation by Lahiru Nuwan Senarathna Wijayasingha Smart Glasses as Support for Training First Responders Abstract:   Link Lab 211

Smart Glasses as Support for Training First Responders

Abstract:  

First responders have to handle many emergency medical incidents every day. The quality of training they undergo is a deciding factor on their performance in real situations. Smart wearable devices can be used to make inferences on the surroundings of a person and draw conclusions on what the wearer is doing. My thesis is on using wearable cameras, microphones and inertial measurement devices to evaluate the training performance of trainee first responders and provide them with real time feedback with hopes to help them improve their training experience and ultimately their performance in the real world focusing on cardiac arrest procedure. This research will be addressing gaps in the existing techniques in order to solve this task. We will be collecting a dataset of participants performing CPR (a critical part of medical procedures related to cardiac arrest situations) and developing benchmark algorithms to evaluate the CPR performance. Measuring depth to objects in a scene is an important task in evaluating various medical procedures. We will develop techniques to measure depth with RBG cameras that can help in situations like measuring the compression depth of CPR procedure and measuring distance to various objects from the wearer of the camera. First responders will manipulate various objects when performing various medical procedures (e.g. using a syringe or a pressure cuff). It is important to measure the amount of deformations that these objects undergo in order to evaluate the quality of the action that the trainee is performing (e.g. to measure the amount of a drug that a syringe has extracted). We will develop techniques to quantify the amount of these deformations in a human readable manner. All of these outputs from our models will be used to provide feedback to the trainee.

 

Committee:

  • Lu Feng, Commitee Chair, CS/ESE/SEAS/UVA
  • John Stankovic, Advisor, CS/SEAS/UVA
  • Homa Alemzadeh, ECE/CS/SEAS/UVA 
  • Bradford Campbell, CS/SEAS/UVA
  • Seongkook Heo, CS/SEAS/UVA