Cinthia Obladen de Almendra Freitas

HANDWRITTEN CHARACTER RECOGNITION USING NONSYMMETRICAL PERCEPTUAL ZONING

In this paper we present an alternative strategy to define zoning for handwriting recognition, which is based on nonsymmetrical perceptual zoning. The idea is to extract some knowledge from the confusion matrices in order to make the zoning process less empirical. The feature set considered in this work is based on concavities/convexities deficiencies, which are obtained by labeling the background pixels of the input image. To better assess the nonsymmetrical zoning we carried out experiments using four different zonings strategies. Experiments show that the nonsymmetrical zoning could be considered as a tool to build more reliable handwriting recognition systems.

Reconstructing strip-shredded documents using color as feature matching

This paper discusses the destroyed documents that have been strip-shredded, which is a often problem in forensic science. The proposed method first extracts features based on color of the boundaries and then computes the nearest neighbor algorithm to carry out the local reconstruction. In this way the overall complexity can be dramatically reduced because few features are used to perform the matching. The preliminary results reported in this paper, which take into account a two hundred documents database, demonstrate that color-matching-based method produces interesting results for the problem of document reconstruction and can be of interest to the forensic document examiners and provide some effective solutions for law enforcement practitioners.

Methodology for the design of NN-based month-word recognizers written on Brazilian bank checks

The study of handwritten words is tied to the development of recognition methods to be used in real-world applications involving handwritten words, such as bank checks, postal envelopes, and handwritten texts, among others. In this work, the focus is handwritten words in the context of Brazilian bank checks, specifically the months of the year, and no restrictions are placed on the types or styles of writing or the number of writers. A global feature set and two architectures of artificial neural networks (ANN) are evaluated for classification of the words. The objectives are to evaluate the performance of conventional and class-modular multiple-layer perceptron (MLP) architectures, to develop a rejection mechanism based on multiple thresholds, and to analyze the behavior of the feature set proposed in the two architectures. The experimental results demonstrate the superiority of the class-modular architecture over the conventional MLP architecture. A rejection mechanism with multiple thresholds demonstrates favorable performance in both architectures. The feature set analysis shows the importance of the structural primitives such as concavities and convexities, and perceptual primitives such as ascenders and descenders. The experimental results reveal a recognition rate of 81.75% without the rejection mechanism, and a reliability rate 91.52% with a rejection rate of 25.33%.

Handwritten Brazilian month recognition: an analysis of two NN architectures and a rejection mechanism

This paper evaluates the use of the conventional architecture feedforward MLP (multiple layer perceptron) and class-modular for the handwriting recognition (HWR) and it also compares the results obtained with previous works in terms of recognition rate. This work presents a feature set in full detail to work with HWR. The experiments showed that the class-modular architecture is better than conventional architecture. The obtained average recognition rates were 77.08% using the conventional architecture and 81.75% using the class-modular. This paper also describes a performance study in which a rejection mechanism with multiple thresholds is evaluated for both conventional and class-modular architectures. The multiple thresholds idea is based on the use of N class-related reject thresholds (CRTs). The results indicate that this rejection mechanism can be used appropriately in both architectures. The experimental results are 86.38% and 91.52% using a handwritten months word database.

Confusion matrix disagreement for multiple classifiers

We present a methodology to analyze Multiple Classifiers Systems (MCS) performance, using the disagreement concept. The goal is to define an alternative approach to the conventional recognition rate criterion, which usually requires an exhaustive combination search. This approach defines a Distance-based Disagreement (DbD) measure using an Euclidean distance computed between confusion matrices and a soft-correlation rule to indicate the most likely candidates to the best classifiers ensemble. As case study, we apply this strategy to two different handwritten recognition systems. Experimental results indicate that the method proposed can be used as a low-cost alternative to conventional approaches.

Zoning and metaclasses for character recognition

The contribution of this paper is twofold. First we investigate the use of the confusion matrices in order to get some insight to better define perceptual zoning for character recognition. The features considered in this work are based on concavities/convexities deficiencies, which are obtained by labeling the background pixels of the input image. Four different perceptual zoning (symmetrical and non-symmetrical) are discussed. Experiments show that this mechanism of zoning could be considered as a reasonable alternative to exhaustive search algorithms. The second contribution is a methodology to define metaclasses for the problem of handwritten character recognition. The proposed approach is based on the disagreement among the characters and it uses Euclidean distance computed between the confusion matrices. Through comprehensive experiments we demonstrate that the use of metaclasses can improve the performance of the system.

Handwritten isolated word recognition: an approach based on Mutual Information for feature set validation

The paper presents the application of the Mutual Information criterion (T.M. Cover and J.A. Thomas, 1991) to validate feature sets extracted from handwritten words in Brazilian legal amounts. The lexicon includes a subset of short words without ascenders/descenders and subsets of words with the same prefix or suffix. These particularities of the Brazilian lexicon show that it is necessary to improve the perpetual feature set with complementary geometric features, and also modeling the prefix and suffix of the words. Finally, the experiments show the viability of our approach.

Crime scene classification

In this paper we provide a study about crime scenes and its features used in criminal investigations. We argue that the crime scene provides a large set of features that can be used to corroborate the conclusions emitted by the experts. We also propose a set of features to classify the violent crime considering two classes: attack from inside or outside of the scene. The classification stage is based on conventional MLP (Multiple-Layer Perceptron) Neural Network and SVM (Support Vector Machine). The experimental results reveal an error rate of 30.3% (MLP), 22.8% (SVM-linear), and 19.4% (SVM-polynomial) using a database composed of 400 crime scenes.

STUDY OF PERCEPTUAL SIMILARITY BETWEEN DIFFERENT LEXICONS

The study investigates the perceptual feature similarity between different lexicons based on visual perception of the words and their representation through an observation sequence. We confirm that it is possible to use databases, which are similar in terms of morphological/perceptual features to improve the recognition performance. In this work, we demonstrated through experimentation, that it is possible to improve the recognition rate of handwritten Portuguese words by adding samples of French words in the training set. Experimental results show the efficiency of this strategy reducing the error rate.

Evaluating the conventional and class-modular architectures feedforward neural network for handwritten word recognition

We evaluate the use of the conventional architecture feedforward MLP (multiple layer perception) and class-modular for the handwriting recognition and it also compares the results obtained with previous works in terms of recognition rate. We present a feature set in full detail to work with handwriting recognition. The experiments showed that the class-modular architecture is better than conventional architecture. The obtained average recognition rates were 77.08% using the conventional architecture and 81.75% using the class-modular.