An improved algorithm of rat sperm image segmentation and counting

摘要

Aiming at the poor real-time characteristics of traditional Ostu for rat sperm image segmentation, we improved this Ostu-based segmentation to reduce its computational time by using Newton iteration. This improved algorithm applies the shape factor s to remove impurities and morphology to mark each sperm with digital tags. The experiments show that this method provides fast and accurate sperm image segmentation and automatic counting of the number of sperms in the images.