A Deep-Learning-Based Approach for in Vitro Maturation Monitoring Utilizing Hexagonal Micropillar Trap Microfluidic Platform

Oocyte maturation is a critical determinant of outcomes in assisted reproductive technologies (ART). Precise control and real-time monitoring of oocyte maturation at the single-cell level are essential for optimizing in vitro maturation (IVM) protocols and enhancing treatment efficacy. Recent advancements in microfluidics have facilitated the creation of controlled microscale culture environments on-chip, enabling tunable nutrient exchange, physiologically relevant shear stress, and cellular interactions. This study proposes a novel micro fluidic platform incorporating a hexagonal micropillar trap (HMT) structure to enable single oocyte trapping and controlled maturation. Integrating a deep-learning-based approach, the system analyzes 24-hour timelapse images to assess cumulus cell area expansion, a key indicator of oocyte maturation. In comparison to conventional culture methods, this platform enables automated oocyte handling, reduces variability, and supports objective and intelligent assessment of the maturation process. The combination of microfluidics and artificial intelligence establishes a robust framework for refining IVM protocols, improving oocyte maturation, and enhancing ART outcomes.