Efficient Bioprocessing: The Potential of Automated microcarrier counting

Researchers have achieved a method for automated microcarrier counting that reduces processing time from minutes to seconds while maintaining a 1.2% median error rate. This task was previously a significant hurdle because manual counting is inherently subjective and prone to human fatigue.

Advancing Automated microcarrier counting in Bioprocessing

Adherent cell culture requires precise quantification to move from mass-based reporting to sophisticated surface-area-dependent control. Current workflows are slowed by manual labour and the limitations of semi-automated software. This early-stage research, currently awaiting peer review, proposes a framework using the Gradient Hough Transform integrated with a Bayesian Optimisation loop. The system's performance was compared against manual expert benchmarks and standard ImageJ plugins:

• Achieved a 1.2% Absolute Percentage Error for counts in aggregate images.
• Estimated average diameters with a 1.7% error margin.
• Outperformed ImageJ in noisy cellular environments, 24.7% error versus 47.2%.

These preliminary results suggest that automated characterisation could standardise production in vaccine and regenerative medicine centres. However, the study does not solve the challenge of accurately identifying critical boundaries within extremely noisy cellular images. While the framework may eventually enable real-time monitoring, these findings remain early-stage until formal validation occurs.