Emmanuel Schmitt

Improving Fuzzy Rule Classifier by Extracting Suitable Features From Capacities With Respect to the Choquet Integral

In this paper, an iterative method to select suitable features in an industrial pattern recognition context is proposed. It combines a global method of feature selection and a fuzzy linguistic rule classifier. It is applied to an industrial fabric textile context. The aim of the global vision system is to identify textile fabric defects. From the related industrial process, the training data sets are small, and some are incomplete. Moreover, the recognition step must be compatible with the time constant of the system, which generally imposes low complexity for the system. The choice of the most relevant features and the reduction of their number are important to respect these constraints. The feature selection method is based on the analysis of indexes extracted on the lattice defined from training in relation with the Choquet integral. This selection step is embedded in an iterative algorithm to discard weaker features in order to decrease the number of rules while keeping good recognition rates. The recognition step is done with a fuzzy reasoning classifier that is well adapted for this application case. The proposed method is quite efficient with small learning data sets because of the generalization capacity of both the feature selection and recognition steps. The experimental study shows the wanted behavior of this approach: the feature number decreases, whereas the recognition rate increases. Thus, the total number of generated fuzzy rules is reduced.

Fuzzy rule classifier: Capability for generalization in wood color recognition

In this paper, a classification method based on fuzzy linguistic rules is exposed. It is applied for the recognition of the gradual color of wood in an industrial context. The wood, which is a natural material, implies uncertainty in the definition of its color. Moreover, the timber context leads obtaining imprecise data. Several factors can have an impact on the sensors (ageing of the acquisition system, variation of the ambient temperature, etc.). Finally, the data sets are often small and incomplete. Thus the proposed method must work within these constraints, and must be compatible with the time-constraint of the system. This generally imposes a weak complexity of the recognition system. The Fuzzy Rule Classifier is split in two main parts, the fuzzification step and the rule generation step. To improve the tuning of this classifier, a specific fuzzification method is presented and compared with more classical ones. Several comparisons have been made with other classification method such as neural network or support vector machine. This experimental study showed the suitability of the proposed approach essentially in term of generalization capabilities from small data sets, and recognition rate improvement.

A Fuzzy Reasoning Classification Method for Pattern Recognition

This paper focuses on a Fuzzy Reasoning Classification Method to improve the potential of pattern recognition in the automated inspection and classification of wooden boards. After the definition of the characteristic features, we implement a fuzzy inference mechanism allowing to take into account the subjectivity of the human visual system. In this article, we have decided to work on the distribution and the representation of the Information. In this sense, our study speaks about the impact of fuzzification on the recognition rates and the structure of our decision module. The part concerning the classification mechanism allows ourselves to integrate knowledge in the generation of the numeric model. This knowledge is enquired at the different field experts thanks to the NIAM formalism. The results, which are presented on a generic benchmark and real data, show the efficiency of such an approach.

Self-Fuzzification Method according to Typicality Correlation for Classification on tiny Data Sets

This article presents a self-fuzzification method to enhance the settings of a fuzzy reasoning classification adapted to the automated inspection of wooden boards. The supervised classification is made thanks to fuzzy linguistic rules generated from small training data sets. This study especially answers to a double industrial need about the pattern recognition in wooden boards. Firstly, few samples are available to generate the recognition model. This aspect makes lesser efficient compilation methods like neural networks in terms of recognition rates. Secondly, the settings of the classification method must be simplified, because the users are not experts in fuzzy logic. In this article, two points are presented. The first part demonstrates the generalization capability of the presented classification method in comparison to more classical algorithms. In the second part, we propose a new automatic method of parameter fuzzification, by using the typicality correlation coefficients of each class.

Fuzzy Rule Iterative Feature Selection (FRIFS) with Respect to the Choquet Integral Apply to Fabric Defect Recognition

An iterative method to select suitable features in an industrial fabric defect recognition context is proposed in this paper. It combines a global feature selection method based on the Choquet integral and a fuzzy linguistic rule classifier. The experimental study shows the wanted behaviour of this approach: the feature number decreases whereas the recognition rate increases. Thus, the number of generated fuzzy rules is reduced.

AN APPEARANCE FUZZY SENSOR INTEGRATING A KNOWLEDGE MODEL

Abstract This article exposes a modeling process of experts knowledge to improve a system of wooden board appearance classification. The aesthetic criteria for the classification are the color and the texture. The extraction of knowledge concerning these two notions is realized with the Natural language Information Analysis Method (NIAM). Then, to improve the current industrial system, we suggest to use this symbolic knowledge model to generate a numeric model. This numeric part is built thanks to a Fuzzy Rules based Inference System (SIF). Fuzzy Sets Theory is here well adapted in order to obtain not-disjointed result classes and to manipulate subjective or symbolic information. Finally, we propose to create our appearance sensor under the form of a “fuzzy sensor”.

Color Recognition Enhancement by Fuzzy Merging

This paper deals with color matching in a wood quality control problem. The main difficulty consists in the recognition of gradual color in an industrial context. The wood, which is a natural material, implies a subjective processing to make the controls. The current methods do not take into account the human aspect of the process. An improvement consists in integrating the imprecision of this subjectivity by using the concept of Fuzzy Sensor. Such a sensor has been developed and done with a Fuzzy Rule Classifier which is quite efficient with imprecise data. Then, in multi-face color matching case, the color recognition is enhanced by merging the outputs of the sensors used together. A specific fuzzy merging operator is proposed to use and compared with more classical merging methods. The obtained results show the efficiency of the proposed enhancement.

Evaluation of adaptive FRIFS method through several classification comparisons

An iterative method to select suitable features for pattern recognition context has been proposed (FRIFS). It combines a global feature selection method based on the Choquet integral and a fuzzy linguistic rule classifier. In this paper, enhancements of this method are presented. An automatic step has been added to make it adaptive to process numerous features. The experimental study, made in a wood defect recognition context, is based on several classifier result analysis. They show the relevancy of the remaining set of selected features. The recognition rates are also considered for each class separately, showing the good behavior of the proposed method.

Fuzzy linguistic rules classifier for wooden board color sorting

This article exposes wood pieces classification method according to their color. The main difficulties encountered by the Company are primarily in the color recognition according to a certain graduality, and the decision to take on all the board with the different sides. These problems imply the use of flexible/robust model and the use of an “intelligent” information management delivered by the sensors. In order to improve the current system, we propose to integrate a method, whose principle is a fuzzy inference system, itself built thanks to fuzzy linguistic rules. The results obtained with our method show a real improvement of the recognition rate compared to a bayesian classifier already used by the Company.