Rapid Colorimetric and Artificial Intelligence- Based Methods for Determining the Microbial Quality of Raw Milk, Processed Milk, and Milk Products

Abstract: Milk is a nutrient-rich food that can support microbial growth when hygienic handling, storage, or processing conditions are inadequate. Conventional microbiological tests remain important reference methods, but they can require substantial time, laboratory resources, and skilled handling. This review examines rapid colorimetric approaches and artificial intelligence (AI)-based methods for screening microbial quality in raw milk, processed milk, and milk products. Particular attention is given to dye-reduction assays such as resazurin, objective RGB colour measurement, machine-learning classification, and sensor systems such as electronic noses. Published studies indicate that colourimetric sensing can convert biochemical activity into measurable optical changes, while machine learning can associate sensor patterns with reference microbial measurements. One recent study using an RGB colourimetric resazurin assay reported 100% prediction accuracy for a low-microbial-concentration class and 96% for a high- microbial-concentration class, with lower performance in the intermediate class. Another study combining electronic- nose measurements with an artificial neural network reported very strong agreement between predicted and reference total bacterial counts in its tested and validated subsets. These findings support the potential of integrated sensing and AI as rapid screening tools, while also showing the need for representative datasets, controlled measurement conditions, external validation, and explainable decision systems. The proposed framework combines rapid colour sensing, reference microbiology, machine learning, and quality classification to support faster dairy-quality monitoring.

Keywords: milk quality, microbial quality, colorimetric assay, resazurin, RGB sensing, artificial intelligence, machine learning, electronic nose, dairy safety