A computer-based behavior monitoring system for low-birth weight infants under intensive care

Abstract

For the intensive care of low-birth weight infants, it is important to continuously monitor behavioral changes. From this point of view, we have developed a computer-based monitoring system to analyze the body movements of low-birth weight infants under intensive care. The system consisted of a video recorder with a micro CCD camera for observing infant's body movements and a real time image processor for analyzing video frames. A 16 bit personal computer equipped with an image processor converted a video frame into a binary image that is a 256 by 256 matrix of pixels. The computer assessed each pixel of the matrix to determine whether the brightness changed between successive binary images or not. The number of changed pixels was counted every 1/30 sec, and such change reflected a infant's body movements. Using the system, body movements of a low-birth weight infant in an incubator were videotaped during 24 hours and analyzed. Results indicate that the behavior monitoring system developed in the present study has not only the ability to detect occurrences of movements of low-birth weight infants, but also the ability to qualitatively describe the pattern of movements.