JoVE - Journal of Visualized Experiments
JoVE Visualize
Contact Us

Robust Real-Time Sperm Tracking with Identity Reassignment Using Extended Kalman Filtering

  • 0Department of Mechanical Engineering, School of Engineering, University of Birmingham, Birmingham B15 2TT, UK.
Sensors (basel, Switzerland) +

|

December 31, 2025

|

December 31, 2025

View abstract on PubMed

Summary

This summary is machine-generated.

This study introduces an enhanced sperm tracking system using an Extended Kalman Filter (EKF) with BoT-SORT. The new method significantly improves sperm identity continuity in microscopy, crucial for fertility diagnostics and Intracytoplasmic Sperm Injection (ICSI).

Area Of Science

  • Biomedical Engineering
  • Computer Vision
  • Reproductive Biology

Background

  • Accurate sperm tracking is vital for automated Intracytoplasmic Sperm Injection (ICSI) and fertility diagnostics.
  • Challenges include identity fragmentation, overcounting, and tracking instability in crowded, low-contrast microscopy.
  • Maintaining correct sperm identities across video frames is crucial for reliable sperm selection.

Purpose Of The Study

  • To develop a robust two-layer tracking framework to enhance sperm identity continuity.
  • To improve the reliability of sperm tracking in challenging microscopy conditions.
  • To integrate Extended Kalman Filter (EKF) with BoT-SORT for superior tracking performance.

Main Methods

  • A two-layer tracking framework integrating BoT-SORT with an Extended Kalman Filter (EKF).
  • EKF models sperm trajectories using position, velocity, and heading for motion prediction and ID correction.
  • Evaluation on the VISEM dataset using standard multi-object tracking (MOT) metrics and trajectory statistics.

Main Results

  • The EKF-BoT-SORT framework improved IDF1 score from 80.30% to 84.84% compared to BoT-SORT.
  • Reduced ID switches from 176 to 132 and decreased ID overcount from 68.75% to 37.5%.
  • Increased average track duration from 74.4 to 91.3 frames, demonstrating enhanced identity preservation.

Conclusions

  • The EKF layer significantly enhances sperm identity preservation in tracking systems.
  • The proposed method maintains real-time feasibility, crucial for clinical applications.
  • Offers a practical foundation for integrating computer vision into ICSI workflows and sperm motility analysis.

Keywords: