Robert Sabourin

An HMM-based approach for off-line unconstrained handwritten word modeling and recognition

Describes a hidden Markov model-based approach designed to recognize off-line unconstrained handwritten words for large vocabularies. After preprocessing, a word image is segmented into letters or pseudoletters and represented by two feature sequences of equal length, each consisting of an alternating sequence of shape-symbols and segmentation-symbols, which are both explicitly modeled. The word model is made up of the concatenation of appropriate letter models consisting of elementary HMMs and an HMM-based interpolation technique is used to optimally combine the two feature sets. Two rejection mechanisms are considered depending on whether or not the word image is guaranteed to belong to the lexicon. Experiments carried out on real-life data show that the proposed approach can be successfully used for handwritten word recognition.

Automatic recognition of handwritten numerical strings: a recognition and verification strategy

A modular system to recognize handwritten numerical strings is proposed. It uses a segmentation-based recognition approach and a recognition and verification strategy. The approach combines the outputs from different levels such as segmentation, recognition, and postprocessing in a probabilistic model. A new verification scheme which contains two verifiers to deal with the problems of oversegmentation and undersegmentation is presented. A new feature set is also introduced to feed the oversegmentation verifier. A postprocessor based on a deterministic automaton is used and the global decision module makes an accept/reject decision. Finally, experimental results on two databases are presented: numerical amounts on Brazilian bank checks and NIST SD19. The latter aims at validating the concept of modular system and showing the robustness of the system using a well-known database.

From dynamic classifier selection to dynamic ensemble selection

In handwritten pattern recognition, the multiple classifier system has been shown to be useful for improving recognition rates. One of the most important tasks in optimizing a multiple classifier system is to select a group of adequate classifiers, known as an Ensemble of Classifiers (EoC), from a pool of classifiers. Static selection schemes select an EoC for all test patterns, and dynamic selection schemes select different classifiers for different test patterns. Nevertheless, it has been shown that traditional dynamic selection performs no better than static selection. We propose four new dynamic selection schemes which explore the properties of the oracle concept. Our results suggest that the proposed schemes, using the majority voting rule for combining classifiers, perform better than the static selection method.

Off-line signature verification using HMM for random, simple and skilled forgeries

The problem of signature verification is in theory a pattern recognition task used to discriminate two classes, original and forgery signatures. Even after many efforts in order to develop new verification techniques for static signature verification, the influence of the forgery types has not been extensively studied. This paper reports the contribution to signature verification considering different forgery types in an HMM framework. The experiments have shown that the error rates of the simple and random forgery signatures are very closed. This reflects the real applications in which the simple forgeries represent the principal fraudulent case. In addition, the experiments show promising results in skilled forgery verification by using simple static and pseudodynamic features.

A comparison of SVM and HMM classifiers in the off-line signature verification

The SVM is a new classification technique in the field of statistical learning theory which has been applied with success in pattern recognition applications like face and speaker recognition, while the HMM has been found to be a powerful statistical technique which is applied to handwriting recognition and signature verification. This paper reports on a comparison of the two classifiers in off-line signature verification. For this purpose, an appropriate learning and testing protocol was created to observe the capability of the classifiers to absorb intrapersonal variability and highlight interpersonal similarity using random, simple and simulated forgeries.

Large vocabulary off-line handwriting recognition: A survey

Considerable progress has been made in handwriting recognition technology over the last few years. Thus far, handwriting recognition systems have been limited to small and medium vocabulary applications, since most of them often rely on a lexicon during the recognition process. The capability of dealing with large lexicons, however, opens up many more applications. This article will discuss the methods and principles that have been proposed to handle large vocabularies and identify the key issues affecting their future deployment. To illustrate some of the points raised, a large vocabulary off-line handwritten word recognition system will be described.

Off-line signature verification by local granulometric size distributions

A fundamental problem in the field of off-line signature verification is the lack of a signature representation based on shape descriptors and pertinent features. The main difficulty lies in the local variability of the writing trace of the signature which is closely related to the identity of human beings. In this paper, we propose a new formalism for signature representation based on visual perception. A signature image consists of 512/spl times/128 pixels and is centered on a grid of rectangular retinas which are excited by local portions of the signature. Granulometric size distributions are used for the definition of local shape descriptors in an attempt to characterize the amount of signal activity exciting each retina on the focus of the attention grid. Experimental evaluation of this scheme is made using a signature database of 800 genuine signatures from 20 individuals. Two types of classifiers, a nearest neighbor and a threshold classifier, show a total error rate below 0.02 percent and 1.0 percent, respectively, in the context of random forgeries.

A METHODOLOGY FOR FEATURE SELECTION USING MULTIOBJECTIVE GENETIC ALGORITHMS FOR HANDWRITTEN DIGIT STRING RECOGNITION

In this paper a methodology for feature selection for the handwritten digit string recognition is proposed. Its novelty lies in the use of a multiobjective genetic algorithm where sensitivity analysis and neural network are employed to allow the use of a representative database to evaluate fitness and the use of a validation database to identify the subsets of selected features that provide a good generalization. Some advantages of this approach include the ability to accommodate multiple criteria such as number of features and accuracy of the classifier, as well as the capacity to deal with huge databases in order to adequately represent the pattern recognition problem. Comprehensive experiments on the NIST SD19 demonstrate the feasibility of the proposed methodology.

Optimization of HVAC Control System Strategy Using Two-Objective Genetic Algorithm

Intelligent building technology for building operation, called the optimization process, is developed and validated in this paper. The optimization process using a multi-objective genetic algorithm will permit the optimal operation of the buildings mechanical systems when installed in parallel with a buildings central control system. Using this proposed optimization process, the supervisory control strategy setpoints, such as supply air temperature, supply duct static pressure, chilled water supply temperature, minimum outdoor ventilation, reheat (or zone supply air temperature), and zone air temperatures are optimized with respect to energy use and thermal comfort. HVAC system steady-state models developed and validated against the monitored data of the existing VAV system are used for energy use and thermal comfort calculations. The proposed optimization process is validated on an existing VAV system for two summer months. Many control strategies applied in a multi-zone HVAC system are also tested and evaluated for one summer day.

A dynamic overproduce-and-choose strategy for the selection of classifier ensembles

The overproduce-and-choose strategy, which is divided into the overproduction and selection phases, has traditionally focused on finding the most accurate subset of classifiers at the selection phase, and using it to predict the class of all the samples in the test data set. It is therefore, a static classifier ensemble selection strategy. In this paper, we propose a dynamic overproduce-and-choose strategy which combines optimization and dynamic selection in a two-level selection phase to allow the selection of the most confident subset of classifiers to label each test sample individually. The optimization level is intended to generate a population of highly accurate candidate classifier ensembles, while the dynamic selection level applies measures of confidence to reveal the candidate ensemble with the highest degree of confidence in the current decision. Experimental results conducted to compare the proposed method to a static overproduce-and-choose strategy and a classical dynamic classifier selection approach demonstrate that our method outperforms both these selection-based methods, and is also more efficient in terms of performance than combining the decisions of all classifiers in the initial pool.