Venous Reservoir Blood Volume Sensor Based on Computer Vision for Clinical Cardiopulmonary Bypass
Ryan Kaddis, MS Computer Science Candidate, CoAS; Alec Lipinski, BS Biomedical Engineering Candidate, CoE; Chan-Jin Chung, PhD, Professor CoAS; Sean Murtha, CCP, Orrum Clinical Analytics, Inc.; Hao Jiang, PhD, Associate Professor, CoE
Computer Science
Lawrence Technological University
submitted by cj424
During a clinical cardiopulmonary bypass (CPB) procedure using a heart-lung machine, the volume of blood inside the venous reservoir is a crucial piece of information for the perfusionist (operator) to effectively manage the patient's blood volume and to make clinical decisions. In existing CPBs, venous reservoir blood volume can only be visually read by the perfusionists from the volume labels on the reservoirs. Such manual reading approaches are generally inconvenient and inefficient. The reading has poor resolution and it is impractical to automatically track the blood volume over a long period. In this poster, we introduce an automated venous reservoir blood volume sensing system for CPBs, leveraging computer vision technology. A camera connected to a computer continuously monitors the blood level within the reservoir in real time, using custom-developed algorithms and software. The system demonstrates high accuracy, with a mean absolute percentage error as low as 2.5%. Extensive testing in a simulated clinical environment confirms that the system is non-intrusive to the blood, does not obstruct operator workflow, and is compatible with various reservoir models.
CC BY
Back to Search