Hyperspectral imaging and deep learning for detection and quantification of germination in Bacillus cereus spores

Abstract

SummaryGermination of Bacillus cereus spores followed by growth and replication of the vegetative cells in food can result in food poisoning and therefore significant economic and health impacts. This study explores a novel approach to detect and differentiate spores and germinated B. cereus cells using hyperspectral imaging (HSI) in combination with machine learning using three different germination triggers. HSI could successfully differentiate between dormant spores, germinated cells and structural controls (non‐spores). The spectral data in the visible‐near‐infrared range are sensitive to unique structural and chemical characteristics specific to spores, setting them apart from their vegetative counterparts and non‐biological controls. This non‐destructive and robust approach shows significant potential for detection and assessment of the physiological state (dormant or germinated). Therefore, HSI is a potential method for the detection of germination in B. cereus spores and merits further research and validation.

Advancing the digital frontier inagri‐foodsupply chains

“…The WPQAS, grounded in machine-learning algorithms, aims to tackle the deceptive practices challenging the wine industry, offering a lightweight and scalable solution. In the realm of microbiological threats, the study by the team of researchers led by Soni et al (2024) pioneers the use of hyperspectral imaging (HSI) and deep learning for the detection and quantification of germination in Bacillus cereus spores. This innovative approach showcases significant potential for non-destructive detection, a crucial advancement in preventing food poisoning.…”