Automated colony detection in fluorescent images using U-Net models

The manual analysis of bacterial colonies in microbiology is time-consuming and prone to errors. Automated solutions can assist scientific staff and enable fully automated processes. The suitability of U-Net models for automatic colony detection was evaluated for fluorescence-based applications. Through targeted hyperparameter settings, U-Net models were trained that demonstrated high segmentation accuracy with average DSC values of 0.941. A particular advantage of this method is the pixel-precise segmentation, which allows for further analyses such as size and distance of colonies to other colonies. The models operated efficiently with a mask prediction time of just 0.27 seconds and required no pre-processing of the images, even in the presence of artifacts. The implementation is based on freely available libraries, making it more accessible. Overall, U-Net models exhibit high potential for automated colony detection and can significantly enhance the efficiency of microbiological processes.