Automatic sperms counting using adaptive local threshold and ellipse detection

A male fertility is influenced by several factors. One of which is the quality of the sperm. Over the last 50 years, human sperm quality decreases. It became the driving force for a male to examine his sperm. It is therefore necessary to have a method to check the sperm quality which is not only accurate, but fast and cheap. Sperm quality can be measured by, one of which, sperm count. In some medical laboratories sperm count is still using manual assessment which makes it subjective, non repetitive, inaccurate, and time consuming. By using multimedia microscope that can record the sperm into a video, one can process it using image processing and computer vision method. In this paper, we propose a method to detect and count sperm in a video taken by a multimedia microscope using an adaptive local threshold and enhanced ellipse detection. The result is 90.97% accuracy compared to the manual assessment.