A Continuous Monitoring Software of the Vision Based Incentive Spirometer: Experimental Validation

Abstract

A continuous observation is required to check as to whether the patient uses the incentive spirometer (IS) properly, and it is difficult to quantify the extent of its use. This study aimed to develop a software to recognize the IS and to check as to whether it is possible to evaluate if the patient uses the IS properly through this software. A flow-oriented IS consisted of the three colored floating balls with inspiratory flows of 700 cc/s, 1200 cc/s, and 1900 cc/s was used. The software was consisted of the IS and its components were recognized in real time from the video recorded by the webcam. It was prepared to recognize yellow, red, and blue, which were the ball colors of the spirometer, and green, which represented the start and end lines of the ball. When the balls and green markers were recognized, the height of the IS was measured. Furthermore, the speed of inhalation was calculated as the ratio of the moving distance of the ball. Lastly, this speed was recorded every 0.02 s. The 500, 1000, and 1500 cc syringes were used to simulate inhalation. After pulling the syringes, the movements of the balls were identified, and the volume of inhaled air was calculated. The calculated volume was statistically compared to the actual volume. When the movement of the ball recognition line in the video was compared with the last moving ball, the software exactly tracked each of the three balls of the IS, and identified as to whether the IS was used and the extent of inhalation. However, it did not exactly reflect the actual inhalation volume. We developed a vision-based monitoring software to measure whether the patients used the IS well and experimentally verified it. The software can be used as a contactless modality for preventing respiratory complications after surgery.