Single sperm tracking using Intersect Cortical Model-Mean Shift Method

Single sperm detection and tracking has received increasing attentions since In Vitro Fertilization (IVF) and Intracytoplasmic Sperm Injection (ICSI) techniques were introduced. In this paper, we proposed an automated system to extract and track single sperm movement. Intersect Cortical Model (ICM) which is derived from Pulse Coupled Neural Network (PCNN) is employed to extract region coordinates and the centroid of the sperm head region. By using the extracted region and centroid as an automated initialization in the proposed method, mean shift based tracking algorithm is then used to track the sperm for the entire video. As a comparison, the proposed method Intersect Cortical Model-Mean shift (ICMMS) has been evaluated with conventional mean shift based tracking method. From the results, the proposed ICMMS method remedies the drawback of mean shift based tracking algorithm by providing more accurate and robust tracking results. After testing with 100 sperm images, less misdetection of sperm has been observed from the results produced by the proposed ICMMS method. In future, the proposed method is expected to be implemented in analyzing male infertility.