An interdisciplinary faculty team at Lawrence Technological University is using computer vision to automate part of the heart-lung machine that stands in for a patient’s heart and lungs during open-heart surgery and other complex medical procedures.
The industry-sponsored project is funded by Orrum Clinical Analytics of Plymouth, LTU’s key partner in establishing and operating the University’s Master of Science in Cardiovascular Perfusion program. Orrum’s CEO, Sean Murtha, a certified clinical perfusionist, brought the project to LTU, hoping to solve a clinical problem: “A perfusionist operates a heart-lung machine to circulate the patient’s blood through an extracorporeal circuit during a cardiac surgery,” Murtha said. “The blood volume inside the venous reservoir is crucial information for the procedure; however, the perfusionists have to manually read the volume labels on the reservoir to measure the volume because there is no automatic blood volume sensor available. Such a manual approach is inefficient, inconvenient, and prone to human errors, which can have direct negative impacts on the patient’s health conditions.”
Chan-Jin (CJ) Chung, Professor of Computer Science, College of Arts and Sciences
Alec Lipinski, Bachelor of Science in Biomedical Engineering candidate, College of Engineering
Ryan Kaddis, Master of Science in Computer Science candidate, College of Arts and Sciences
Hao Jiang, Associate Professor of Biomedical Engineering, College of Engineering
So, Hao Jiang, assistant professor of electrical and computer engineering at LTU, got to work. “Starting in fall 2021, we began developing a blood volume sensor for this application, funded by Orrum,” Jiang said. “To measure blood in the venous reservoir is not as easy as measuring coffee in a coffee maker. Firstly, the sensor cannot have any direct contact with blood because any contamination of the patient’s blood is strictly prohibited. Secondly, the sensor cannot obstruct a perfusionist’s view of the reservoir or block their access to it. Thirdly, the sensor should work for multiple brands of reservoirs and should be convenient for perfusionists to deploy. By summer 2023, our development efforts focused on sensing mechanisms based on contact image sensors and gravimetric sensors. Our works were very fruitful, and we published one journal article and filed a patent application in 2023. However, we also identified limitations for their applicability in clinical use because they may block the perfusionists’ view to the reservoirs.”
To tackle this challenge, Jiang said, “Starting in the fall of 2023, we assembled an interdisciplinary team comprised of biomedical engineers, computer scientists, and perfusionists to develop a new blood volume sensor based on computer vision algorithm. One key advantage of using computer vision is that a compact, low-cost web camera can be conveniently positioned towards the reservoir, and the blood volume can be automatically read by image processing. This scheme allows the sensor to automatically work for all major brands of reservoirs, accomplishing the optimal versatility.”
This work was recently published in the Institute of Electrical and Electrical Engineers’ peer-reviewed journal IEEE Sensors Letters, and a provisional patent application has been filed.
Jiang credited CJ Chung, professor of computer science at LTU, and his student, Ryan Kaddis, for adding critical computer vision expertise to the project. Said Jiang: “They brought their top-class expertise to this project and have been the main workforce innovating the detection algorithm. As a faculty member in biomedical engineering, we’ve been incredibly fortunate on the BME side to collaborate with such outstanding computer science folks on this project. This is a fantastic example of interdisciplinary collaboration at its best.”
Jiang said the project has so far provided training to nine BME undergraduate students, including two BME senior design teams, as well as someone Jiang said was “a very inventive BME student who joined us recently, Alec Lipinski. This project also trained one student in computer science. These students contributed to the publication of two journal articles and two patent applications. We are so proud of our team’s productivity!”
And, Jiang said, the team is “deeply grateful to have been collaborating with perfusionists throughout our development process. Their clinical insights have been invaluable in shaping the technical design of our sensors. We appreciate the funding support and guidance from Sean Murtha. We want to thank Robb Johnson from Orrum for helping us test our sensors in their SIMLAB. We also want to thank the support from BME faculty and staff.”
Finally, Chung added: “Computer vision is a field within computer science that enables computers to ‘see’ and interpret images and videos. This project exemplifies interdisciplinary research, demonstrating how computer vision can enhance existing cardiovascular perfusion systems through automatic reading and logging of the current blood volume level in the reservoir. After completing the CV system, I am particularly interested in developing machine learning models to predict the optimal blood volume level in real time for open-heart surgery patients within specific patient groups.”
And, Chung said, “I would like to thank Dr. Jiang and the Biomedical Engineering Department for initiating this project and providing an exciting research and development opportunity for computer science.”
[1] Tanner Foley, Alex Fernández-Rajal i Sabala, Makenna Fockler, Zahra Alzayer, Sean Murtha, and Hao Jiang. “Continuous contactless measurement of blood volume inside venous reservoirs for cardiopulmonary bypass.” IEEE Sensors Journal 23 (2023): 14882-14890. https://doi.org/10.1109/JSEN.2023.3275963
[2] Sean Murtha, Hao Jiang, Tanner Foley, Alex Fernández-Rajal I Sabala, Makenna Fockler, Zahra Alzayer, Ahron Wayne. “Blood Volume Sensor System.” PCT Patent Application WO2023107591A1, published on Jun 15, 2023. [Patent pending]
[3] Ryan Kaddis, Chan-Jin Chung, Sean Murtha and Hao Jiang, “A Non-intrusive, Non-obstructive, Versatile Venous Reservoir Blood Volume Sensor Based on Computer Vision for Clinical Cardiopulmonary Bypass,” IEEE Sensors Letters. (published online on Mar 17, 2025) https://doi.org/10.1109/LSENS.2025.3551948
[4] Sean Murtha, Ryan Kaddis, Chan-Jin Chung, and Hao Jiang. “Venous reservoir blood volume sensor with computer vision”, U.S. Provisional Patent Application No. 63/707,913, filed on Oct 16, 2024. [Patent pending]
By: Matt Roush
