期刊名称：IAENG International Journal of Computer Science
印刷版ISSN：1819-656X
电子版ISSN：1819-9224
出版年度：2018
卷号：45
期号：4
页码：531-539
出版社：IAENG - International Association of Engineers
摘要：One of the determinants of sperm quality is themotility of spermatozoa. The motility of spermatozoa ismeasured by microscopic sperm test. Conventionally; thedetermination of sperm motility is performed by experts, inwhich the judgment tends to be subjective. The existence ofComputer-Assisted Sperm Analysis (CASA) is beneficial insolving problems related to the emergence of subjectivity in thedetermination of sperm motility. Generally, CASA andresearchers in this field use phase contrast microscopes toobtain images with higher contrast. In this study, the positionand motility determinations of spermatozoa in the video wereperformed using video records taken from a bright fieldmicroscope with low contrast, along with various otherdeficiencies. With a combination of several stages of works,namely frame difference background subtraction, contrastsettingwith Otsu threshold as an indicator, filtering processusing mathematical morphology to determine the position ofobjects, as well as linear regression and root mean square value(RMS) calculations. From the results of experimental testsperformed on human spermatozoa video data, the abovemethod indicated that the positions of sperm motility fromtracking results had recognizable trajectories based on theaverage distance position to the linear regression line, with anRMS threshold of 10. There were ten progressive spermatozoaand four non-progressive spermatozoa. The method usedsuccessfully determined 14 human spermatozoa. There were71% progressive spermatozoa, while the remaining 29% werenon-progressive. Under the WHO 2010 guidelines, a 71%percentage indicates normal sperm motility.
关键词：Spermatozoa; Linear Regression; Root Mean;Square; Background Subtraction; Mathematical Morphology
