Image analysis as support for sensory evaluation of food

Analysis of images is widely used as a rapid method in different steps of food production and quality control such as fruit and vegetable selection, monitoring of production processes, control of fruit ripening and food shelf life, just to cite few examples. Another context where image analysis provides important support is the correlation between human evaluation and sensory analysis in order to improve several aspects of food evaluation. In the present contribution, as for example of applications in the context of sensory analysis, we report two case studies were a visual analyser (IRIS VA 300, AlphaMos), a camera-based imaging system designed for visual assessment of products appearance, was employed. The first study deals with the grading of Fuji and Gala apples based on skin stripes patterning that is an important commercial parameter and is used to reward producers. The acquired images were used to compare different multivariate models for classification of the apples into the 9 grading categories defined by human evaluators. In the second study, we show how the analysis of images can support the quality control of Trentingrana cheeses that foresees the systematic sensory quality evaluation of production wheels performed by a panel of experts. Based on this control the associated dairies receive price premiums or penalties. The sensory evaluation protocol is based on seven quality parameters, four of which are related to visual aspects: “exterior aspect of the wheel”, “rind thickness”, “texture” and “interior colour”. This study aims at the development of predicting models based of images acquired with a visual analyser for the measurement of parameters related to visual aspects that can assist in the objective evaluation of sensory quality attributes of Trentingrana. Both studies show how optimization of predictive models based on instrumental measurements can support the human sensory assessment of food in a routine sensory evaluation.