Industry Sponsored
BME
2025-2026
Fall
Winter
Spring
PartiClear
Summary
Technical Approach/Methodology
PartiClear is built to make particulate testing more efficient, consistent, and scalable for medical device production. Manual titanium particulate inspection can be slow and operator-dependent, creating challenges for high-volume quality testing.
By integrating a motorized stage controller with machine learning-based particle detection, PartiClear helps automate particulate counting and measurement while supporting reliable pass/fail reporting. Beyond improving workflow efficiency, this project reflects the role of engineering in protecting human welfare by strengthening the quality control processes behind devices used in delicate surgical procedures.
Outcomes
Automated Slide Scanning
The stage controller moves the filter slide under the microscope in a controlled snake-path pattern, reducing the need for manual slide positioning during particulate inspection.
Machine Learning Detection
A YOLO-based detection model identifies titanium particulates in microscope images and distinguishes them from other debris such as dirt or lint.
Particle Counting & Measurement
PartiClear uses detection bounding boxes to count titanium particles and estimate particle size, helping support more standardized particulate analysis.
Pass/Fail Reporting
The system is designed to generate a final report summarizing particulate counts by size threshold and identifying whether slides pass or fail quality testing requirements.