Neural network classification: a cork industry case

In this paper we investigate an Intelligent Computer Vision System applied for recognition and classification of commercially available tiles from the cork industry. The system is capable of acquiring and processing gray images using several feature generation and analysis techniques. Its functionality includes image acquisition, feature recognition and extraction, data preprocessing (Analysis of Variance and Principal Component Analysis) and feature classification with neural networks (NN). The system is investigated in terms of statistical feature processing (number of features and dimensionality reduction techniques) and classifier design (NN architecture, topology, target coding, and complexity of training). We report system test and validation results of the recognition and classification tasks with up to 95% success rate. Some of those results are due to our investigation and combination of feature generation techniques: application of Analysis of Variance (ANOVA) and Principal Component Analysis (PCA), which appeared to be very efficient for preprocessing the data; and use of suitable NN design and learning method. The employed NNs are trained with our genetic low-discrepancy search method (GLPTS) for global optimisation. The obtained and reported results demonstrate strongly competitive nature when compared with results from other authors investigating similar systems.