[Show abstract][Hide abstract]ABSTRACT: We present a new method to detect the presence of the hollow heart, an internal disorder of the potato tubers, using hyperspectral imaging technology in the infrared region. A set of 468 hyperspectral
cubes of images has been acquired from Agria variety potatoes, that have been cut later to check the presence of a hollow
heart. We developed several experiments to recognize hollow heart potatoes using different Artificial Intelligence and Image
Processing techniques. The results show that Support Vector Machines (SVM) achieve an accuracy of 89.1% of correct classification.
This is an automatic and non-destructive approach, and it could be integrated into other machine vision developments.
[Show abstract][Hide abstract]ABSTRACT: We present a benchmarking framework to design multi spectral systems working in the NIR range for multiple purposes. This framework is composed of a hyperspectral imaging hardware and an ad-hoc software that performs pattern recognition experiments (image acquisition, segmentation, feature extraction, feature selection, classification and evaluation steps) comparing different algorithms in every step. For each experiment, we obtain a solution using a generic hyper spectral system, but we also obtain enough data to design a specific multi-spectral system in order to decrease the overall execution time. This improvement is based in the feature se lection step, that provides the most relevant wavelengths for the problem. The framework has been tested for detecting internal and external features in potatoes, determining the origin of honey, and studying fecundity parameters in hen eggs.
[Show abstract][Hide abstract]ABSTRACT: The study of biology and population dynamics of fish species requires the estimation of fecundity parameters in individual
fish in many fisheries laboratories. The traditional procedure used in fisheries research is to classify and count the oocytes
manually on a subsample of known weight of the ovary, and to measure few oocytes under a binocular microscope. With an adequate
interactive tool, this process might be done on a computer. However, in both cases the task is very time consuming, with the
obvious consequence that fecundity studies are not conducted routinely. In this work we develop a computer vision system for
the classification of oocytes using texture features in histological images. The system is structured in three stages: 1)
extraction of the oocyte from the original image; 2) calculation of a texture feature vector for each oocyte; and 3) classification
of the oocytes using this feature vector. A statistical evaluation of the proposed system is presented and discussed.
[Show abstract][Hide abstract]ABSTRACT: The common scab is a skin disease of the potato tubers that decreases the quality of the product and influences significantly the price. We present an objective and non-destructive method to detect the common scab on potato tubers using an experimental hyperspectral imaging system. A supervised pattern recognition experiment has been performed in order to select the best subset of bands and classification algorithm for the problem. Support Vector Machines (SVM) and Random Forest classifiers have been used. We map the amount of common scab in a potato tuber by classifying each pixel in its hyperspectral cube. The result is the percentage of the surface affected by common scab. Our system achieves a 97.1% of accuracy with the SVM classifier.
[Show abstract][Hide abstract]ABSTRACT: This paper has a double objective. The first goal was to develop an authentication system to differentiate between traditional orujo alcoholic distillates with and without a certified brand of origin (CBO). Owing to their low price and quality, samples without a CBO can be used as substrates for falsification of genuine CBO ones. The second objective was to perform a comparison of the abilities of the different chemometric procedures employed for this classification. The classification was performed on the basis of the chemical information contained in the metal composition of the orujo distillates. Eight metals determined by electrothermal atomic absorption spectrometry and inductively coupled plasma optical emission spectrometry were considered (Ca, Cd, Cr, Cu, K, Mg, Na and Ni). After the appropriate pretreatment, the data were processed using different chemometric techniques. In the first stage, principal component analysis and cluster analysis were employed to reveal the latent structure contained in the data. Once it had been demonstrated that a relation exists between the metal composition and the raw materials, and not between the metal composition and the distillation systems employed for the orujo production, the second step consisted in the comparative application of different supervised pattern recognition procedures (such as linear discriminant analysis, K-nearest neighbours, soft independent modelling of class analogy, UNEQ and different artificial neural network approaches, including multilayer feed-forward, support vector machines, learning vector quantization and probabilistic neural networks). The results showed the different capabilities of the diverse classification techniques to discriminate between Galician orujo samples. The best results were those provided by probabilistic neural networks, in which the correct recognition abilities for CBO classes and without CBO classes were 98.6 +/- 3.1 and 98.0 +/- 4.5%; the prediction results were 87.7 +/- 3.3 and 86.2 +/- 5.0%, respectively. The usefulness of chemical metal analysis in combination with chemometric techniques to develop a classification procedure to authenticate Galician CBO orujo samples is demonstrated.
Article · Jul 2010 · Analytical and Bioanalytical Chemistry
[Show abstract][Hide abstract]ABSTRACT: This paper proposes polytope ARTMAP (PTAM), an adaptive resonance theory (ART) network for classification tasks which does not use the vigilance parameter. This feature is due to the geometry of categories in PTAM, which are irregular polytopes whose borders approximate the borders among the output predictions. During training, the categories expand only towards the input pattern without category overlap. The category expansion in PTAM is naturally limited by the other categories, and not by the category size, so the vigilance is not necessary. PTAM works in a fully automatic way for pattern classification tasks, without any parameter tuning, so it is easier to employ for nonexpert users than other classifiers. PTAM achieves lower error than the leading ART networks on a complete collection of benchmark data sets, except for noisy data, without any parameter optimization.
Full-text Article · Oct 2007 · IEEE Transactions on Neural Networks
[Show abstract][Hide abstract]ABSTRACT: PolyTope ARTMAP (PTAM)  is an ART neural network based on internal categories with irregular polytope (polygon in IR
) geometry. Categories in PTAM do not overlap, so that their expansion is limited by the other categories, and not by the
category size. This makes the vigilance parameter unnecessary. What happens if categories have irregular geometries but overlap
is allowed? This paper presents Overlapping PTAM (OPTAM), an alternative to PTAM based on polytope overlapping categories,
which tries to answer this question. The comparison between the two approaches in classification tasks shows that category
overlap does not reduce neither the classification error nor the number of categories, and it also requires vigilance as a
tuning parameter. Futhermore, OPTAM provides a significant variability in the results among different data sets.
[Show abstract][Hide abstract]ABSTRACT: The present work describes an improved version of MART (Multichannel ART), a neural network aimed at the adaptive recognition of multichannel patterns. As is habitual in ART networks, MART is directed
at problems that require unsupervised learning, but it has a greater level of adaptability to the characteristics of the input
patterns, selectively evaluating the different signal channels on which it operates and the classes learnt, modulating the
discrimination capacity, and dynamically learning/forgetting classes with regard to their level of representativity.
[Show abstract][Hide abstract]ABSTRACT: This paper describes MART, an ART-based neural network for adaptive classification of multichannel signal pat- terns without prior supervised learning. Like other ART-based classifiers, MART is especially suitable for situations in which not even the number of pattern categories to be distinguished is known a priori; its novelty lies in its truly multichannel orientation, especially its ability to quantify and take into account during pattern classification the different changing reliabilities of the individual signal channels. The extent to which this ability can reduce the creation of spurious or duplicate categories (a major problem for ART-based classifiers of noisy signals) is illustrated by evaluation of its performance in classifying QRS complexes in two-channel ECG traces which were taken from the MIT-BIH database and contaminated with noise.
[Show abstract][Hide abstract]ABSTRACT: The use of time frequency distributions (TFDs) with adaptive kernels for the spectral estimation of non stationary signals has been shown to be an extremely useful tool in many applications. Nevertheless, their high computational cost, due to the necessity to calculate a new kernel in each time instance, poses an important problem in real time applications. Given that many real signals show intervals of relative stationarity, in this work we propose an algorithm for the control of the instant of adaptation in this type of technique, based on the detection of situations of quasi-stationarity of the signal. By way of the analysis of real and artificial signals, and applying our algorithm to a specific TFD, we clearly show the advantages of this new technique.
[Show abstract][Hide abstract]ABSTRACT: In this paper we present an algorithm that performs reasoning processes on fuzzy knowledge bases with chaining between rules. The algorithm we propose permits obtaining valid inferences even in situations where part of the variables in the knowledge base are unknown. The support for the description of the execution algorithm is provided by the Petri net formalism, which organizes in a convenient way all the information in the base, and expresses in a simple way this and other processes performed onto it. This simplicity is mainly achieved also because the execution of the fuzzy knowledge base is carried out in a parameterized truth space using the linguistic truth values defined by J.F. Baldwin (1979), which reduces the computational cost of the process and makes easier its mapping onto the Petri net formalism