Team of UMaine Students place second at the Northeast Bioengineering Conference

UMaine students Jacob Holbrook, Dominic Kugell, Margaret McCarthy, and Basel White have had a busy semester. The team of undergraduate students presented their senior capstone project at the Northeast Bioengineering Conference (NEBEC), where they placed second overall out of the 44 undergraduate poster presentations.

Supervised by Professor of Practice Robert Bowie, and mentored by Professor of Biomedical Engineering Andre Khalil, the students also participated in the 2022 UMaine Student Symposium, a week before their presentation at the NEBEC event. The conference took place in person at Columbia University in New York City on April 23 and 24.

Their project titled “Non-Contact Image Analysis of Breathing Rate using an Unmanned Aerial Vehicle (UAV)” examines how unmanned aerial vehicles or UAVs, commonly known as drones, can be used to gather vital life signs in search and rescue missions.

Land-based search and rescue missions have limitations and safety concerns, especially where helicopters are required to navigate the terrain. Drones, however, are much smaller and easier to operate compared to conventional aircrafts, allowing search and rescue teams the ability to collect information on a patient’s condition and better prepare the team to treat the patient. Using UAVs allows search and rescue teams to be more accurate and safe when performing missions.

In the field, a drone needs to land 20 feet away from the patient, per Federal Aviation Administration guidelines, and then takes videos of the patient to assess their current state. These videos are sent back to the UAV operator for assessment.

The drones can also be used to fly in medication and equipment like life jackets or splints to the patient to keep them stable until a ground team can reach them. However, currently there is no effective way to assess a patient’s vital signs using UAV technology. To address this problem, the group set out to find a way to assess a patient’s breathing pattern.

The team used two seperate image processing techniques to calculate breathing rate: image thresholding and sliding-window convolution. Image thresholding is a type of image segmentation, where users can change the pixels in an image to make them easier to analyze. Sliding-window convolution is another technique to analyze images which looks at a particular window in each image and looks for changes from window to window.

The team is able to get accurate breathing patterns and rates using both methodologies, and was even able to replicate the patterns at lower video resolution. This means that search and rescue teams would be able to decrease the video processing time without losing information.

Through their research, the team has successfully shown that they can land a UAV 20 feet from a patient, analyze a 15 second video, and produce accurate breathing patterns and rates for a patient.

This is helpful to search and rescue teams and medical staff who are preparing to rescue and treat a patient. They have rapid access to vital signs and insight into the condition of the patient. In a medical emergency where time is of the essence, this could serve as a life saving resource.

In the field, this technique would be applied using user-friendly software that allows search and rescue teams to upload video, select the area they would like to analyze, and receive graphs with breathing rate.