Sperm Motility Analysis by using Recursive Kalman Filters with the smartphone based data acquisition and reporting approach

Ilhan H. O., Yüzkat M., Aydin N.

EXPERT SYSTEMS WITH APPLICATIONS, vol.186, 2021 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 186
  • Publication Date: 2021
  • Doi Number: 10.1016/j.eswa.2021.115774
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Aerospace Database, Applied Science & Technology Source, Communication Abstracts, Computer & Applied Sciences, INSPEC, Metadex, Public Affairs Index, Civil Engineering Abstracts
  • Keywords: Sperm Motility Analysis, Trajectory clustering, Video stabilization, Recursive Kalman Filter tracking, Videomicroscopy analysis, Biomedical video processing, ACROSOME INTEGRITY, FRAMEWORK, TRACKING, TEXTURE, IMAGES
  • Yıldız Technical University Affiliated: Yes


Semen analysis is currently performed by using two techniques. Visual assessment technique is manual observation based technique and strongly depends on the experiences of the observer. Therefore, the reliability of the results is skeptical. On the other hand, computer based expert systems are more consistent and reliable. However, they are very expensive systems, therefore, cannot be utilized in many laboratories. In this study, we proposed a hybrid expert system utilizing visual assessment environment with the computerized analyzing part to eliminate the disadvantages of each technique. In the proposed system, smartphone based data acquisition approach is used to provide more modular and practical expert system for the sperm analysis. The records are, then, transferred to the server to analyze by developed software. In this analyzing software, we proposed multistage hybrid analyzing approach in terms of video stabilization, sperm concentration and motility analysis. Each video was initially fixed by the Speed Up Robust Features based matching technique. Then, Kalman Filter was employed for sperm tracking. After tracking step, trajectories have been divided into 3 s length to prevent possible incorrect assignments due to sudden changes in sperm motions. In the experimental tests, we combined all trajectories obtained from a total of 18 videos of 6 different subjects. We clustered a total of 89438 trajectories into 4 cluster as fast progressive, progressive, non-progressive and immotile according to extracted seven features. In order to compare the results, we also analyzed the same semen sample in another expert system, SQA-Vision. The difference was measured 3.4% and 4.8% in the determination of total and motile sperm concentration, and 2.1%, 7.4%, 5.3% for progressive, non-progressive and immotile movement type analysis respectively. The significance and impact of the proposed system are capability of reporting more detailed results in a variety of situations and having more advantages than any expert systems utilized for sperm analysis in terms of portability, cost and modularity. Additionally, to the best of our knowledge, this is the first study reporting use of the smartphone in an expert system for the sperm analysis in terms of data acquisition and result reporting.