Predicting freshness of shell eggs using a technique based on the dieletric properties

Given the non-invasive and non-destructive nature, the measure- ment of the dielectric properties of agri-food products is the object of several studies. Dielectric characteristics of the material vary with moisture content, density, composition and structure, water activ- ity and can be measured with techniques which range from direct current to microwaves (Nelson 1991). Recently, a simple dielectric technique based on a resonant plate capacitor probe was set up to predict the freshness of shell eggs (Ragni et al., 2006). Inspired from biological nervous systems, artii cial neural networks (ANN) is a mathematical algorithm capable of relate input and out- put parameters by learning from example trough iterations. ANN is able to approximate any non-linear input output relationship by means of a simple structure with connections in parallel between neurons. h is computing method is widely applied in food quality evaluation especially using computer vision system (Goyache et al., 2001; Du and Sun, 2004; Du and Sun, 2006). Main applications con- cern the classii cation of cereal grains (Lou et al., 1999), fruits and vegetables (Brandon et al., 1990; Kavdir and Guyer, 2002) and meat (Chandraratne et al., 2007). h e present research attempts to evaluate the days of storage and the main freshness parameters of shell eggs on the basis of non- destructive spectroscopic dielectric analysis and artii cial neural network method. In conclusion, even if preliminary results, this work shows that dielectric spectroscopy could be used to set up a non-destructively technique able to predict the freshness of the shell eggs and it opens up new perspectives for on line applications.