Automated design of a computer vision system for visual food quality evaluation. http://dx.doi.org/10.1007/s11947-012-0934-2

Abstract

Considerable research eorts in computer vision applied to food quality evaluation have been developed in the last years, however, they have been concentrated on using or developing tailored methods based on visual features that are able to solve a specic task.Nevertheless, today’s computer capabilities are giving us new ways to solve complex computer vision problems. In particular, a new paradigm based on machine learningtechniques has emerged posing the task of recognizing visual patterns as a search problem based on training data and a hypothesis space composed by visual features and suitable classiers. Furthermore, now we are able to extract, process, and test in thesame time more image features and classiers than before. Thus, we propose a general framework that designs a computer vision system automatically, i.e., it nds {without human interaction{ the features and the classiers for a given application avoiding the classical trial and error framework commonly used by human designers. The key idea of the proposed framework is to select from a large set of features and a bank of classi-ers those features and classiers that achieve the highest performance. We tested our framework on eight dierent food quality evaluation problems yielding a classication performance of 95% or more in every case. The proposed framework was implemented as a Matlab Toolbox available for non commercial purposes.Keywords: Food quality evaluation; feature extraction; feature selection; classication; pattern recognition; image analysis; image segmentation; image processing; computer vision.