Identification of Abnormal Spermatozoa Motility Using the SVM Algorithm

Karim, Mohammad Daniel Sulthonul
Puspaningrum, Eva Yulia
Diyasa, I Gede Susrama Mas

Literasi Nusantara
Vol. 5 No. 1 (2025): Literasi Nusantara: November 2024- February 2025

Spermatozoa motility is one of the key indicators in determining male fertility quality. Manual assessment of motility abnormalities often requires significant time and effort, thus necessitating a more efficient and accurate automated approach. This study aims to identify abnormalities in spermatozoa motility using the Support Vector Machine (SVM) algorithm, utilizing microscopic video data analyzed through TrackPy for spermatozoa trajectory tracking. The analysis process involves data acquisition, spermatozoa detection in each frame, sperm trajectory construction, and trajectory classification into normal or abnormal categories. The SVM model was trained using a dataset derived from spermatozoa trajectories classified based on parameters such as average velocity and trajectory linearity. The results show that the method achieved the highest accuracy of 89 percent in identifying spermatozoa motility abnormalities in HD resolution videos with a frame rate of 30 fps.