Food microstructure in the data science era: From images to values

Quantifying microstructure in foods holds the key to understanding how food materials are hierarchically structured at different length scales. This review highlights the efforts towards quantitative analysis of food microstructure and aims at facilitating the field of food microstructures to move towards a quantitative regime. Consensus on classifications of food microstructure is currently still missing and can therefore lead to multiple interpretations. In addition, the microstructures of foods are widely varied and typically not as clear or distinct as in other disciplines. Data science approaches, including machine learning and deep learning, are not new to the field of image analysis; they are, however, only starting to be used for segmentation and quantification of food microstructure. Overall, the shift in focus to a more quantitative science will in turn move the field towards the rational design of foods, including the tailored design of textural and other structure-related attributes of food.