High resolution acquisition and processing of fluorescent microscopic images for quality assurance of food

• Novel approach on investigations of bacterial contamination by fluorescent microscopy.
• An innovative fluorescent microscope in inverse setup was developed.
• It consists of an image acquisition system and a 3D positioning system.
• Machine learning technique is effectually used for automated germ analysis.
• Segmentation, feature extraction and classification results are presented.

This paper contents a novel approach on investigations of bacterial contamination by fluorescent microscopy. Incipiently presenting the motivation and state-of-the-art methods of germ analysis, the challenges and solutions in hardware construction design are pointed out. The fluorescent microscope developed in inverse setup consists necessarily of an image acquisition system and a positioning system in three degrees of freedom. Afterwards a description of automated germ analysis evaluated with machine learning technique is followed. The segmentation method, feature extraction and several classifiers are described in detail concluded in a broad range of applications.