Muriel Visani | Université De La Rochelle (original) (raw)

Papers by Muriel Visani

Dans ce papier, nous nous intéressons à la structure de treillis de Galois, treillis utilisé dans... more Dans ce papier, nous nous intéressons à la structure de treillis de Galois, treillis utilisé dans la méthode Navigala, méthode de reconnaissance de symboles basée sur un parcours (de type arbre de décision) dans le treillis, et plus généralement aux treillis dits treillis dichotomiques, définis à partir d'attributs binaires issus d'un traitement de discrétisation. Nous mettons en évidence les liens structurels unissant les arbres de décision et les treillis dichotomiques en montrant tout d'abord que tout arbre de décision est inclus dans le treillis, mais également que le treillis est en fait la fusion de tous les arbres de décision. Nous finissons par des expérimentations visant à comparer, pour de la reconnaissance de symboles, les performances des arbres de classification et des treillis construits avec la méthode Navigala.

In this paper, we propose a new approach for symbol recognition using structural signatures and a... more In this paper, we propose a new approach for symbol recognition using structural signatures and a Ga- lois Lattice as classifier. The structural signatures are based on topological graphs computed from segments which are extracted from the symbol images by using an adapted Hough transform. These structural sig- natures, that can be seen as dynamic paths which carry high level

The first competition on music scores that was organized at ICDAR in 2011 awoke the interest of r... more The first competition on music scores that was organized at ICDAR in 2011 awoke the interest of researchers, who participated both at staff removal and writer identification tasks. In this second edition, we focus on the staff removal task and simulate a real case scenario: old music scores. For this purpose, we have generated a new set of images using two kinds of degradations: local noise and 3D distortions. This paper describes the dataset, distortion methods, evaluation metrics, the participant’s methods and the obtained results.

The first competition on music scores that was organized at ICDAR and GREC in 2011 awoke the inte... more The first competition on music scores that was organized at ICDAR and GREC in 2011 awoke the interest of researchers, who participated both at staff removal and writer identification tasks. In this second edition, we propose a staff removal competition where we simulate old music scores. Thus, we have created a new set of images, which contain noise and 3D distortions. This paper describes the distortion methods, metrics, the participant’s methods and the obtained results.

Kanungo noise model is widely used to test the robustness of different binary document image anal... more Kanungo noise model is widely used to test the robustness of different binary document image analysis methods towards noise. This model only works with binary images while most document images are in grayscale. Because binarizing a document image might degrade its contents and lead to a loss of information, more and more researchers are currently focusing on segmentation-free methods (Angelika et al [2]). Thus, we propose a local noise model for grayscale images. Its main principle is to locally degrade the image in the neighbourhoods of “seed-points” selected close to the character boundary. These points define the center of “noise regions”. The pixel values inside the noise region are modified by a Gaussian random distribution to make the final result more realistic. While Kanungo noise models scanning artifacts, our model simulates degradations due to the age of the document itself and printing/writing process such as ink splotches, white specks or streaks. It is very easy for us...

2010 12th International Conference on Frontiers in Handwriting Recognition, 2010

This paper deals with on-line handwriting recognition in a closed-world environment with a large ... more This paper deals with on-line handwriting recognition in a closed-world environment with a large lexicon. Several applications using handwriting recognition have been developed, but most of them consider a lexicon of limited size. Many difficulties, in particular confusions during the segmentation stage, are linked to the use of a large lexicon, with large writing variations and an increased complexity of

Lecture Notes in Computer Science, 2008

In this paper, we propose a new approach for symbol recognition using structural signatures and a... more In this paper, we propose a new approach for symbol recognition using structural signatures and a Galois Lattice as classifier. The structural signatures are based on topological graphs computed from segments which are extracted from the symbol images by using an adapted Hough transform. These structural signatures, which can be seen as dynamic paths which carry high level information, are robust towards various transformations. They are classified by using a Galois Lattice as a classifier. The performances of the proposed approach are evaluated on the GREC03 symbol database and the experimental results we obtain are encouraging.

Dans cet article, nous rappelons la méthode de classification supervisée Navigala, que nous avons... more Dans cet article, nous rappelons la méthode de classification supervisée Navigala, que nous avons développée pour de la reconnaissance de symboles détériorés. Elle repose sur une navigation dans un treillis de Galois similaireà une navigation dans un arbre de décision. Les treillis manipulés par Navigala sont des treillis dits dichotomiques, dont nous décrivons dans ce papier les propriétés et les liens structurels avec les arbres de décision. La construction du treillis de Galois obligeà uneétape préalable de discrétisation des données continues (discrétisation globale), ce qui n'est généralement pas le cas de l'arbre de décision qui procèdeà cette discrétisation au cours de sa construction (discrétisation locale). Utilisée comme prétraitement, la discrétisation détermine les concepts et la taille du treillis, lorsque l'algorithme de génération est directement appliqué sur ces données discrétisées. Nous proposons donc un algorithme de discrétisation locale pour la construction du treillis dichotomique ce qui pourrait nous permettre de mettre en oeuvre une méthode d'élagage en cours de génération et ainsi d'améliorer les performances du treillis etéviter le sur-apprentissage.

Proceedings of the 2nd International Workshop on Historical Document Imaging and Processing - HIP '13, 2013

ABSTRACT This paper presents an e�cient parametrization method for generating synthetic noise on ... more ABSTRACT This paper presents an e�cient parametrization method for generating synthetic noise on document images. By specify- ing the desired categories and amount of noise, the method is able to generate synthetic document images with most of degradations observed in real document images (ink splotches, white specks or streaks). Thanks to the ability of simulating di�erent amount and kind of noise, it is possible to evaluate the robustness of many document image analysis methods. It also permits to generate data for algorithms that employ a learning process. The degradation model presented in [7] needs eight parameters for generating randomly noise re- gions. We propose here an extension of this model which aims to set automatically the eight parameters to gener- ate precisely what a user wants (amount and category of noise). Our proposition consists of three steps. First, Nsp seed-points (i.e. centres of noise regions) are selected by an adaptive procedure. Then, these seed-points are classi�ed into three categories of noise by using a heuristic rule. Fi- nally, each size of noise region is set using a random process in order to generate degradations as realistic as possible.

Techniques et sciences informatiques, 2010

ABSTRACT Dans cet article nous proposons une nouvelle approche pour la reconnaissance de symboles... more ABSTRACT Dans cet article nous proposons une nouvelle approche pour la reconnaissance de symboles bruités utilisant une signature structurelle et un treillis de Galois comme classifieur. La signature structurelle est basée sur des graphes topologiques calculés à partir de segments extraits des images de symboles par une adaptation de la transformée de Hough aux images de symboles. Les signatures ainsi construites caractérisent les relations entre segments et portent des informations de haut niveau, ce qui leur confère une certaine robustesse vis-à-vis de certaines transformations. Les symboles ainsi caractérisés sont classés en utilisant un treillis de Galois (ou treillis des concepts) comme classifieur, car ce classifieur a montré sa robustesse visà-vis de bruits. Les performances de l'approche proposée ont été évaluées sur la base de symboles de GREC03 et les résultats obtenus sont encourageants, en particulier en ce qui concerne la robustesse de notre méthodevis-à-vis de la présence de bruit.

IJDAR, 2011

Most document analysis applications rely on the extraction of shape descriptors, which may be gro... more Most document analysis applications rely on the extraction of shape descriptors, which may be grouped into different categories, each category having its own advantages and drawbacks (O.R. Terrades et al. in Proceedings of ICDAR'07, pp. 227-231, 2007). In order to improve the richness of their description, many authors choose to combine multiple descriptors. Yet, most of the authors who propose a new descriptor content themselves with comparing its performance to the performance of a set of single state-of-the-art descriptors in a specific applicative context (e.g. symbol recognition, symbol spotting…). This results in a proliferation of the shape descriptors proposed in the literature. In this article, we propose an innovative protocol, the originality of which is to be as independent of the final application as possible and which relies on new quantitative and qualitative measures. We introduce two types of measures: while the measures of the first type are intended to characterize the descriptive power (in terms of uniqueness, distinctiveness and robustness towards noise) of a descriptor, the second type of measures characterizes the complementarity between multiple descriptors. Characterizing upstream the complementarity of shape descriptors is an alternative to the usual approach where the descriptors to be combined are selected by trial and error, considering the performance characteristics of the overall system. To illustrate the contribution of this protocol, we performed experimental studies using a set of descriptors and a set of symbols which are widely used by the community namely ART and SC descriptors and the GREC 2003 database.

Image Analysis and Recognition, 2004

In this paper, a new statistical projection-based method called Two-Dimensional-Oriented Linear D... more In this paper, a new statistical projection-based method called Two-Dimensional-Oriented Linear Discriminant Analysis (2DO-LDA) is presented. While in the Fisherfaces method the 2D image matrices are first transformed into 1D vectors by merging their rows of pixels, 2DO-LDA is directly applied on matrices, as 2D-PCA. Within and between-class image covariance matrices are generalized, and 2DO-LDA aims at finding a projection space jointly maximizing the second and minimizing the first by considering a generalized Fisher criterion defined on image matrices. A series of experiments was performed on various face image databases in order to evaluate and compare the effectiveness and robustness of 2DO-LDA to 2D-PCA and the Fisherfaces method. The experimental results indicate that 2DO-LDA is more efficient than both 2D-PCA and LDA when dealing with variations in lighting conditions, facial expression and head pose.

2012 IEEE RIVF International Conference on Computing & Communication Technologies, Research, Innovation, and Vision for the Future, 2012

In recent years, the expansion of acquisition devices such as digital cameras, the evolution of s... more In recent years, the expansion of acquisition devices such as digital cameras, the evolution of storage and transmission techniques of multimedia documents and the developement of tablet computers facilitate the growing of many large image databases as well as the interactions with the users. This increases the need for efficient and robust methods for finding information in these huge masses of data, including feature extraction methods and feature space structuring methods. The feature extraction methods aim to extract feature descriptors for each image. The feature space structuring methods organize indexed images in order to facilitate, accelerate and improve the results of further retrieval. Clustering is one kind of feature space structuring. Clustering may organize the dataset into groups of similar objects without prior knowledge (unsupervised clustering) or with a limited amount of prior knowledge (semi-supervised clustering). In this article, we present both formal and experimental comparisons of different unsupervised clustering methods for structuring large image databases. We use different image databases of increasing sizes (Wang, PascalVoc2006, Caltech101, Corel30k) to study the scalability of the different approaches. Moreover, a summary of semi-supervised clustering methods is presented and an interactive semi-supervised clustering model using the HMRF-kmeans is experimented on the Wang image database in order to analyse the improvement of the clustering results when user feedbacks are provided.

2013 12th International Conference on Document Analysis and Recognition, 2013

ABSTRACT Unconstrained on-line handwriting recognition is typically approached within the framewo... more ABSTRACT Unconstrained on-line handwriting recognition is typically approached within the framework of generative HMMbased classifiers. In this paper, we introduce a novel discriminative method that relies, in contrast, on explicit grapheme segmentation and SVM-based character recognition. In addition to single character recognition with rejection, bi-characters are recognized in order to refine the recognition hypotheses. In particular, bi-character recognition is able to cope with the problem of shared character parts. Whole word recognition is achieved with an efficient dynamic programming method similar to the Viterbi algorithm. In an experimental evaluation on the Unipen-ICROW-03 database, we demonstrate improvements in recognition accuracy of up to 8% for a lexicon of 20,000 words with the proposed method when compared with an HMM-based baseline system. The computational speed is on par with the baseline system.

2013 12th International Conference on Document Analysis and Recognition, 2013

ABSTRACT This article presents a method for generating semisynthetic images of old documents wher... more ABSTRACT This article presents a method for generating semisynthetic images of old documents where the pages might be torn (not flat). By using only 2D deformation models, most existing methods give non-realistic synthetic document images. Thus, we propose to use 3D approach for reproducing geometric distortions in real documents. First, our new texture coordinate generation technique extracts texture coordinates of each vertex in the document shape (mesh) resulting from 3D scanning of a real degraded document. Then, any 2D document image can be overlayed on the mesh by using an existing texture image mapping method. As a result, many complex real geometric distortions can be integrated in generated synthetic images. These images then can be used for enriching training sets or for performance evaluation. Our degradation method here is jointly used with the character degradation model we proposed in [1] to generate the 6000 semi-synthetic degraded images of the music score removal staff line competition of ICDAR 2013

Proceedings. IEEE Conference on Advanced Video and Signal Based Surveillance, 2005., 2005

In this paper, we present a novel approach for face recognition, using a new subspace method call... more In this paper, we present a novel approach for face recognition, using a new subspace method called bilinear discriminant analysis (BDA) and normalized radial basis function networks (NRBFNs). In a first step, BDA extracts the features that enhance separation between classes by using a generalized bilinear projection-based Fisher criterion, computed from image matrices directly. In a second step, the features

Lecture Notes in Computer Science, 2005

In this paper, a new statistical projection method called Bilinear Discriminant Analysis (BDA) is... more In this paper, a new statistical projection method called Bilinear Discriminant Analysis (BDA) is presented. The proposed method efficiently combines two complementary versions of Two-Dimensional-Oriented Linear Discriminant Analysis (2DoLDA), namely Column-Oriented Linear Discriminant Analysis (CoLDA) and Row-Oriented Linear Discriminant Analysis (RoLDA), through an iterative algorithm using a generalized bilinear projectionbased Fisher criterion. A series of experiments was performed on various international face image databases in order to evaluate and compare the effectiveness of BDA to RoLDA and CoLDA. The experimental results indicate that BDA is more efficient than RoLDA, CoLDA and 2DPCA for the task of face recognition, while leading to a significant dimensionality reduction.

2010 20th International Conference on Pattern Recognition, 2010

This paper deals with on-line handwriting recognition. Analytic approaches have attracted an incr... more This paper deals with on-line handwriting recognition. Analytic approaches have attracted an increasing interest during the last ten years. These approaches rely on a preliminary segmentation stage, which remains one of the most difficult problems and may affect strongly the quality of the global recognition process. In order to circumvent this problem, this paper introduces a bi-character model, where each character is recognized jointly with its neighboring characters. This model yields two main advantages. First, it reduces the number of confusions due to connections between characters during the character recognition step. Second, it avoids some possible confusion at the character recognition level during the word recognition stage. Our experimentation on significant databases shows some interesting improvements of the recognition rate, since the recognition rate is increased from 65% to 83% by using this bi-character strategy.

2011 International Conference on Document Analysis and Recognition, 2011

In this paper, we present two text-independent writer identification methods in a closed-world co... more In this paper, we present two text-independent writer identification methods in a closed-world context. Both methods use on-line and off-line features jointly with a classifier inspired from information retrieval methods. These methods are local, respectively based on the character and grapheme levels. This writer identification engine may be used to personalize our cursive word recognition engine [1] to the handwriting style of the writer, resulting in an adaptive cursive word recognizer. Experiments assess the effectiveness of the proposed approaches in a context of writer identification as well as integrated to our cursive word recognizer to make it adaptive.

Dans ce papier, nous nous intéressons à la structure de treillis de Galois, treillis utilisé dans... more Dans ce papier, nous nous intéressons à la structure de treillis de Galois, treillis utilisé dans la méthode Navigala, méthode de reconnaissance de symboles basée sur un parcours (de type arbre de décision) dans le treillis, et plus généralement aux treillis dits treillis dichotomiques, définis à partir d'attributs binaires issus d'un traitement de discrétisation. Nous mettons en évidence les liens structurels unissant les arbres de décision et les treillis dichotomiques en montrant tout d'abord que tout arbre de décision est inclus dans le treillis, mais également que le treillis est en fait la fusion de tous les arbres de décision. Nous finissons par des expérimentations visant à comparer, pour de la reconnaissance de symboles, les performances des arbres de classification et des treillis construits avec la méthode Navigala.

In this paper, we propose a new approach for symbol recognition using structural signatures and a... more In this paper, we propose a new approach for symbol recognition using structural signatures and a Ga- lois Lattice as classifier. The structural signatures are based on topological graphs computed from segments which are extracted from the symbol images by using an adapted Hough transform. These structural sig- natures, that can be seen as dynamic paths which carry high level

The first competition on music scores that was organized at ICDAR in 2011 awoke the interest of r... more The first competition on music scores that was organized at ICDAR in 2011 awoke the interest of researchers, who participated both at staff removal and writer identification tasks. In this second edition, we focus on the staff removal task and simulate a real case scenario: old music scores. For this purpose, we have generated a new set of images using two kinds of degradations: local noise and 3D distortions. This paper describes the dataset, distortion methods, evaluation metrics, the participant’s methods and the obtained results.

The first competition on music scores that was organized at ICDAR and GREC in 2011 awoke the inte... more The first competition on music scores that was organized at ICDAR and GREC in 2011 awoke the interest of researchers, who participated both at staff removal and writer identification tasks. In this second edition, we propose a staff removal competition where we simulate old music scores. Thus, we have created a new set of images, which contain noise and 3D distortions. This paper describes the distortion methods, metrics, the participant’s methods and the obtained results.

Kanungo noise model is widely used to test the robustness of different binary document image anal... more Kanungo noise model is widely used to test the robustness of different binary document image analysis methods towards noise. This model only works with binary images while most document images are in grayscale. Because binarizing a document image might degrade its contents and lead to a loss of information, more and more researchers are currently focusing on segmentation-free methods (Angelika et al [2]). Thus, we propose a local noise model for grayscale images. Its main principle is to locally degrade the image in the neighbourhoods of “seed-points” selected close to the character boundary. These points define the center of “noise regions”. The pixel values inside the noise region are modified by a Gaussian random distribution to make the final result more realistic. While Kanungo noise models scanning artifacts, our model simulates degradations due to the age of the document itself and printing/writing process such as ink splotches, white specks or streaks. It is very easy for us...

2010 12th International Conference on Frontiers in Handwriting Recognition, 2010

This paper deals with on-line handwriting recognition in a closed-world environment with a large ... more This paper deals with on-line handwriting recognition in a closed-world environment with a large lexicon. Several applications using handwriting recognition have been developed, but most of them consider a lexicon of limited size. Many difficulties, in particular confusions during the segmentation stage, are linked to the use of a large lexicon, with large writing variations and an increased complexity of

Lecture Notes in Computer Science, 2008

In this paper, we propose a new approach for symbol recognition using structural signatures and a... more In this paper, we propose a new approach for symbol recognition using structural signatures and a Galois Lattice as classifier. The structural signatures are based on topological graphs computed from segments which are extracted from the symbol images by using an adapted Hough transform. These structural signatures, which can be seen as dynamic paths which carry high level information, are robust towards various transformations. They are classified by using a Galois Lattice as a classifier. The performances of the proposed approach are evaluated on the GREC03 symbol database and the experimental results we obtain are encouraging.

Dans cet article, nous rappelons la méthode de classification supervisée Navigala, que nous avons... more Dans cet article, nous rappelons la méthode de classification supervisée Navigala, que nous avons développée pour de la reconnaissance de symboles détériorés. Elle repose sur une navigation dans un treillis de Galois similaireà une navigation dans un arbre de décision. Les treillis manipulés par Navigala sont des treillis dits dichotomiques, dont nous décrivons dans ce papier les propriétés et les liens structurels avec les arbres de décision. La construction du treillis de Galois obligeà uneétape préalable de discrétisation des données continues (discrétisation globale), ce qui n'est généralement pas le cas de l'arbre de décision qui procèdeà cette discrétisation au cours de sa construction (discrétisation locale). Utilisée comme prétraitement, la discrétisation détermine les concepts et la taille du treillis, lorsque l'algorithme de génération est directement appliqué sur ces données discrétisées. Nous proposons donc un algorithme de discrétisation locale pour la construction du treillis dichotomique ce qui pourrait nous permettre de mettre en oeuvre une méthode d'élagage en cours de génération et ainsi d'améliorer les performances du treillis etéviter le sur-apprentissage.

Proceedings of the 2nd International Workshop on Historical Document Imaging and Processing - HIP '13, 2013

ABSTRACT This paper presents an e�cient parametrization method for generating synthetic noise on ... more ABSTRACT This paper presents an e�cient parametrization method for generating synthetic noise on document images. By specify- ing the desired categories and amount of noise, the method is able to generate synthetic document images with most of degradations observed in real document images (ink splotches, white specks or streaks). Thanks to the ability of simulating di�erent amount and kind of noise, it is possible to evaluate the robustness of many document image analysis methods. It also permits to generate data for algorithms that employ a learning process. The degradation model presented in [7] needs eight parameters for generating randomly noise re- gions. We propose here an extension of this model which aims to set automatically the eight parameters to gener- ate precisely what a user wants (amount and category of noise). Our proposition consists of three steps. First, Nsp seed-points (i.e. centres of noise regions) are selected by an adaptive procedure. Then, these seed-points are classi�ed into three categories of noise by using a heuristic rule. Fi- nally, each size of noise region is set using a random process in order to generate degradations as realistic as possible.

Techniques et sciences informatiques, 2010

ABSTRACT Dans cet article nous proposons une nouvelle approche pour la reconnaissance de symboles... more ABSTRACT Dans cet article nous proposons une nouvelle approche pour la reconnaissance de symboles bruités utilisant une signature structurelle et un treillis de Galois comme classifieur. La signature structurelle est basée sur des graphes topologiques calculés à partir de segments extraits des images de symboles par une adaptation de la transformée de Hough aux images de symboles. Les signatures ainsi construites caractérisent les relations entre segments et portent des informations de haut niveau, ce qui leur confère une certaine robustesse vis-à-vis de certaines transformations. Les symboles ainsi caractérisés sont classés en utilisant un treillis de Galois (ou treillis des concepts) comme classifieur, car ce classifieur a montré sa robustesse visà-vis de bruits. Les performances de l'approche proposée ont été évaluées sur la base de symboles de GREC03 et les résultats obtenus sont encourageants, en particulier en ce qui concerne la robustesse de notre méthodevis-à-vis de la présence de bruit.

IJDAR, 2011

Most document analysis applications rely on the extraction of shape descriptors, which may be gro... more Most document analysis applications rely on the extraction of shape descriptors, which may be grouped into different categories, each category having its own advantages and drawbacks (O.R. Terrades et al. in Proceedings of ICDAR'07, pp. 227-231, 2007). In order to improve the richness of their description, many authors choose to combine multiple descriptors. Yet, most of the authors who propose a new descriptor content themselves with comparing its performance to the performance of a set of single state-of-the-art descriptors in a specific applicative context (e.g. symbol recognition, symbol spotting…). This results in a proliferation of the shape descriptors proposed in the literature. In this article, we propose an innovative protocol, the originality of which is to be as independent of the final application as possible and which relies on new quantitative and qualitative measures. We introduce two types of measures: while the measures of the first type are intended to characterize the descriptive power (in terms of uniqueness, distinctiveness and robustness towards noise) of a descriptor, the second type of measures characterizes the complementarity between multiple descriptors. Characterizing upstream the complementarity of shape descriptors is an alternative to the usual approach where the descriptors to be combined are selected by trial and error, considering the performance characteristics of the overall system. To illustrate the contribution of this protocol, we performed experimental studies using a set of descriptors and a set of symbols which are widely used by the community namely ART and SC descriptors and the GREC 2003 database.

Image Analysis and Recognition, 2004

In this paper, a new statistical projection-based method called Two-Dimensional-Oriented Linear D... more In this paper, a new statistical projection-based method called Two-Dimensional-Oriented Linear Discriminant Analysis (2DO-LDA) is presented. While in the Fisherfaces method the 2D image matrices are first transformed into 1D vectors by merging their rows of pixels, 2DO-LDA is directly applied on matrices, as 2D-PCA. Within and between-class image covariance matrices are generalized, and 2DO-LDA aims at finding a projection space jointly maximizing the second and minimizing the first by considering a generalized Fisher criterion defined on image matrices. A series of experiments was performed on various face image databases in order to evaluate and compare the effectiveness and robustness of 2DO-LDA to 2D-PCA and the Fisherfaces method. The experimental results indicate that 2DO-LDA is more efficient than both 2D-PCA and LDA when dealing with variations in lighting conditions, facial expression and head pose.

2012 IEEE RIVF International Conference on Computing & Communication Technologies, Research, Innovation, and Vision for the Future, 2012

In recent years, the expansion of acquisition devices such as digital cameras, the evolution of s... more In recent years, the expansion of acquisition devices such as digital cameras, the evolution of storage and transmission techniques of multimedia documents and the developement of tablet computers facilitate the growing of many large image databases as well as the interactions with the users. This increases the need for efficient and robust methods for finding information in these huge masses of data, including feature extraction methods and feature space structuring methods. The feature extraction methods aim to extract feature descriptors for each image. The feature space structuring methods organize indexed images in order to facilitate, accelerate and improve the results of further retrieval. Clustering is one kind of feature space structuring. Clustering may organize the dataset into groups of similar objects without prior knowledge (unsupervised clustering) or with a limited amount of prior knowledge (semi-supervised clustering). In this article, we present both formal and experimental comparisons of different unsupervised clustering methods for structuring large image databases. We use different image databases of increasing sizes (Wang, PascalVoc2006, Caltech101, Corel30k) to study the scalability of the different approaches. Moreover, a summary of semi-supervised clustering methods is presented and an interactive semi-supervised clustering model using the HMRF-kmeans is experimented on the Wang image database in order to analyse the improvement of the clustering results when user feedbacks are provided.

2013 12th International Conference on Document Analysis and Recognition, 2013

ABSTRACT Unconstrained on-line handwriting recognition is typically approached within the framewo... more ABSTRACT Unconstrained on-line handwriting recognition is typically approached within the framework of generative HMMbased classifiers. In this paper, we introduce a novel discriminative method that relies, in contrast, on explicit grapheme segmentation and SVM-based character recognition. In addition to single character recognition with rejection, bi-characters are recognized in order to refine the recognition hypotheses. In particular, bi-character recognition is able to cope with the problem of shared character parts. Whole word recognition is achieved with an efficient dynamic programming method similar to the Viterbi algorithm. In an experimental evaluation on the Unipen-ICROW-03 database, we demonstrate improvements in recognition accuracy of up to 8% for a lexicon of 20,000 words with the proposed method when compared with an HMM-based baseline system. The computational speed is on par with the baseline system.

2013 12th International Conference on Document Analysis and Recognition, 2013

ABSTRACT This article presents a method for generating semisynthetic images of old documents wher... more ABSTRACT This article presents a method for generating semisynthetic images of old documents where the pages might be torn (not flat). By using only 2D deformation models, most existing methods give non-realistic synthetic document images. Thus, we propose to use 3D approach for reproducing geometric distortions in real documents. First, our new texture coordinate generation technique extracts texture coordinates of each vertex in the document shape (mesh) resulting from 3D scanning of a real degraded document. Then, any 2D document image can be overlayed on the mesh by using an existing texture image mapping method. As a result, many complex real geometric distortions can be integrated in generated synthetic images. These images then can be used for enriching training sets or for performance evaluation. Our degradation method here is jointly used with the character degradation model we proposed in [1] to generate the 6000 semi-synthetic degraded images of the music score removal staff line competition of ICDAR 2013

Proceedings. IEEE Conference on Advanced Video and Signal Based Surveillance, 2005., 2005

In this paper, we present a novel approach for face recognition, using a new subspace method call... more In this paper, we present a novel approach for face recognition, using a new subspace method called bilinear discriminant analysis (BDA) and normalized radial basis function networks (NRBFNs). In a first step, BDA extracts the features that enhance separation between classes by using a generalized bilinear projection-based Fisher criterion, computed from image matrices directly. In a second step, the features

Lecture Notes in Computer Science, 2005

In this paper, a new statistical projection method called Bilinear Discriminant Analysis (BDA) is... more In this paper, a new statistical projection method called Bilinear Discriminant Analysis (BDA) is presented. The proposed method efficiently combines two complementary versions of Two-Dimensional-Oriented Linear Discriminant Analysis (2DoLDA), namely Column-Oriented Linear Discriminant Analysis (CoLDA) and Row-Oriented Linear Discriminant Analysis (RoLDA), through an iterative algorithm using a generalized bilinear projectionbased Fisher criterion. A series of experiments was performed on various international face image databases in order to evaluate and compare the effectiveness of BDA to RoLDA and CoLDA. The experimental results indicate that BDA is more efficient than RoLDA, CoLDA and 2DPCA for the task of face recognition, while leading to a significant dimensionality reduction.

2010 20th International Conference on Pattern Recognition, 2010

This paper deals with on-line handwriting recognition. Analytic approaches have attracted an incr... more This paper deals with on-line handwriting recognition. Analytic approaches have attracted an increasing interest during the last ten years. These approaches rely on a preliminary segmentation stage, which remains one of the most difficult problems and may affect strongly the quality of the global recognition process. In order to circumvent this problem, this paper introduces a bi-character model, where each character is recognized jointly with its neighboring characters. This model yields two main advantages. First, it reduces the number of confusions due to connections between characters during the character recognition step. Second, it avoids some possible confusion at the character recognition level during the word recognition stage. Our experimentation on significant databases shows some interesting improvements of the recognition rate, since the recognition rate is increased from 65% to 83% by using this bi-character strategy.

2011 International Conference on Document Analysis and Recognition, 2011

In this paper, we present two text-independent writer identification methods in a closed-world co... more In this paper, we present two text-independent writer identification methods in a closed-world context. Both methods use on-line and off-line features jointly with a classifier inspired from information retrieval methods. These methods are local, respectively based on the character and grapheme levels. This writer identification engine may be used to personalize our cursive word recognition engine [1] to the handwriting style of the writer, resulting in an adaptive cursive word recognizer. Experiments assess the effectiveness of the proposed approaches in a context of writer identification as well as integrated to our cursive word recognizer to make it adaptive.