Per Munkevik, Tom Duckett and Gunnar Hall

Vision System Learning for Ready Meal Characterisation

Proceedings of the International Conference on Engineering and Food (ICEF9)


Abstract

This paper introduces a new method for automatic grading of ready meals using machine vision.The different components of a meal are segmented and features are extracted. A self-organising feature map (SOFM) is then trained using examples of meals graded as "approved" by sensory experts. The trained SOFM is then used to discriminate invalid meals from valid ones, by considering the similarity between the test samples and the acquired sensory model as a measure of food quality.
Key words: ready meal, vision system, neural network, image analysis, sensory values.

Download

[pdf]

Bibtex 

@INPROCEEDINGS{Munkevik:2004,
   AUTHOR =       "Per Munkevik and Tom Duckett and Gunnar Hall",
   TITLE =        "Vision System Learning for Ready Meal Characterisation",
   BOOKTITLE =    "Proc. International Conference on Engineering and Food (ICEF9)",
   YEAR =         "2004",
   PAGES =        "",
   ADDRESS =      "Montpellier, France",
   DATE =         "March, 7--11",
}