C. Viard-gaudin | Université de Nantes (original) (raw)

Papers by C. Viard-gaudin

2008 International Symposium on Information Science and Engineering, 2008

In this paper we experiment the capabilities of Hidden Markov Models (HMM) to model the time-vari... more In this paper we experiment the capabilities of Hidden Markov Models (HMM) to model the time-variant signal produced by the movement of a pen when drawing a sketch such as an electrical circuit diagram. We consider that the sketches have been generated by a two-level stochastic process. The underlying process governs the stroke production from a neuro-motor control point of

SPIE Proceedings, 2011

We propose in this paper a new online handwritten flowchart database and perform some first exper... more We propose in this paper a new online handwritten flowchart database and perform some first experiments to have a baseline benchmark on this dataset. The collected database consists of 78 flowcharts labeled at the stroke and symbol levels. In addition, an isolated database of graphical and text symbols was extracted from these collected flowcharts. Then, we tackle the problem of online handwritten flowchart recognition from two different points of view. Firstly, we consider that flowcharts are correctly segmented, and we propose different classifiers to perform two tasks, text/non-text separation and graphical symbol recognition. Tested with the extracted isolated test database, we achieve up to 99% and 96% in text/non-text separation and up to 81.3% in graphical symbols recognition. Secondly, we propose a global approach to perform flowchart segmentation and recognition. For this latter, we adopt a global learning schema and a recognition architecture that considers a simultaneous segmentation and recognition. Global architecture is trained and tested directly with flowcharts. Results show the interest of such global approach, but regarding the complexity of flowchart segmentation problem, there is still lot of space to improve the global learning and recognition methods.

Proceedings of the 9th International Conference on Neural Information Processing, 2002. ICONIP '02., 2002

Ninth International Conference on Document Analysis and Recognition (ICDAR 2007), 2007

Handwriting is an alternative method for entering texts composing Short Message Services. However... more Handwriting is an alternative method for entering texts composing Short Message Services. However, a whole new language features the texts which are produced. They include for instance abbreviations and other consonantal writing which sprung up for time saving and fashion. We have collected and processed a significant number of such handwriting SMS, and used various strategies to tackle this challenging area of handwriting recognition. We proposed to study more specifically three different phenomena: consonant skeleton, rebus, and phonetic writing. For each of them, we compare the rough results produced by a standard recognition system with those obtained when using a specific language model.

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

Dynamic Time Warping (DTW) is a famous distance to compare two mono-stroke symbols. It obeys the ... more Dynamic Time Warping (DTW) is a famous distance to compare two mono-stroke symbols. It obeys the boundary and continuity constraints. The extension to multi-stroke symbols raises specific problems. A naïve solution is to convert the multi-stroke symbol into a single one by a direct concatenation respecting the handwriting order. However, people may write a symbol with different stroke orders and different stroke directions. Applying a brute force method by searching all the possible directions and orders leads to prohibitive calculation times. To reduce the searching complexity, we propose the DTW-A* algorithm that keeps the continuity constraint during each partial matching. This DTW-A* distance achieves the best recognition rate and the best stability in cross-validation when comparing three distances (DTW-A*, DTW, Modified Hausdorff Distance) on a flowchart dataset which mainly contains multi-stroke symbols.

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

An unconstrained online handwritten Chinese text lines dataset, SCUT-COUCH2009-TL, a subset of SC... more An unconstrained online handwritten Chinese text lines dataset, SCUT-COUCH2009-TL, a subset of SCUT-COUCH [1], is built to facilitate the research of unconstrained online Chinese text recognition. Texts for handcopying are sampled from China Daily corpus with a stratified random manner. The current vision of SCUT-COUCH2009-TL has 8,809 text lines (4,813 lines are collected by touch screen LCD and 3,996 by digital pen) and 159,866 characters in total that are written by more than 157 participants. To demonstrate that the dataset is practical, an over-segmentation, dynamic programming and semantic model based algorithm was presented for segmenting and recognizing the unconstrained online Chinese text lines. In preliminary experiments on the dataset, the proposed algorithm recognition achieves a baseline accuracy of 56.41%.

Document Recognition and Retrieval XIX, 2012

ABSTRACT Online handwritten data, produced with Tablet PCs or digital pens, consists in a sequenc... more ABSTRACT Online handwritten data, produced with Tablet PCs or digital pens, consists in a sequence of points (x, y). As the amount of data available in this form increases, algorithms for retrieval of online data are needed. Word spotting is a common approach used for the retrieval of handwriting. However, from an information retrieval (IR) perspective, word spotting is a primitive keyword based matching and retrieval strategy. We propose a framework for handwriting retrieval where an arbitrary word spotting method is used, and then a manifold ranking algorithm is applied on the initial retrieval scores. Experimental results on a database of more than 2,000 handwritten newswires show that our method can improve the performances of a state-of-the-art word spotting system by more than 10%.

2009 10th International Conference on Document Analysis and Recognition, 2009

Hybrid of Neural Network (NN) and Hidden Markov Model (HMM) has been popular in word recognition,... more Hybrid of Neural Network (NN) and Hidden Markov Model (HMM) has been popular in word recognition, taking advantage of NN discriminative property and HMM representational capability. However, NN does not guarantee good generalization due to Empirical Risk minimization (ERM) principle that it uses. In our work, we focus on online word recognition using the support vector machine (SVM) for character recognition. SVM's use of structural risk minimization (SRM) principle has allowed simultaneous optimization of representational and discriminative capability of the character recognizer. We evaluated SVM in isolated character recognition environment using IRONOFF and UNIPEN character database. We then demonstrate the practical issues in using SVM within a hybrid setting with HMM for word recognition by testing the hybrid system on the IRONOFF word database and obtained commendable results.

Proceedings of IEEE Region 10 International Conference on Electrical and Electronic Technology. TENCON 2001 (Cat. No.01CH37239)

This paper describes an offline cursive handwritten word recognition system that combines Hidden ... more This paper describes an offline cursive handwritten word recognition system that combines Hidden Markov Models (HMM) and Neural Networks (NN). Using a fast left-right slicing method, we generate a segmentation graph that describes all possible ways to segment a word into letters. The NN computes the observation probabilities for each letter hypothesis in the segmentation graph. Then, the HMMs compute likelihood for each word in the lexicon by summing the probabilities over all possible paths through the graph. We present the preprocessing and the recognition process as well as the training procedure for the NN-HMM hybrid system. Another recognition system based on discrete HMMs is also presented for performance comparison. The latter is also used for bootstrapping the NN-HMM hybrid system. Recognition performances of the two recognition systems using two image databases of French isolated words are presented. This paper is one of the first publications using the IRONOFF database, and thus will be used as a reference for future work on this database.

Une alternative a l'emploi de la technologie laser pour la realisation de lecteurs de codes-b... more Une alternative a l'emploi de la technologie laser pour la realisation de lecteurs de codes-barres est presentee dans cet article, la solution proposee est basee sur une approche par traitement d'image. On a developpe pour cela un systeme global comportant deux fonctions principales: la localisation du code-barres dans l'image, puis son decodage prenant en compte l'ensemble des pixels du code-barres. La localisation est realisee grâce a une methode originale d'extraction de regions denses en gradients mono-orientes. La lecture est basee sur la detection des transitions entre barres par extraction des passages par zero de la derivee seconde du profil de projection reconstruit a partir de l'image. Ce systeme a ete evalue sur une base de plusieurs centaines d'images, les resultats obtenus montrent un taux de reussite global (localisation et lecture) de 100 % sur une tres large plage de variation des parametres d'acquisition, notamment la resolution et le...

An offline cursive handwriting recognition system, based on hybrid of Neural Networks (NN) and Hi... more An offline cursive handwriting recognition system, based on hybrid of Neural Networks (NN) and Hidden Markov Models (HMM), is described in this paper. Applying SegRec principle, the recognizer does not make hard decision at the character segmentation process. Instead, it delays the character segmentation to the recognition stage by generating a segmentation graph that describes all possible ways to segment a word into letters. To recognize a word, the NN computes the observation probabilities for each segmentation candidates (SCs) in the segmentation graph. Then, using concatenated letter-HMMs, a likelihood is computed for each word in the lexicon by multiplying the probabilities over the best paths through the graph. We present in detail two approaches to train the word recognizer: 1). character-level training 2). word-level training. The recognition performances of the two systems are discussed. I.

Support vector machine (SVM) is one of the recent methods for statistical learning, it addresses ... more Support vector machine (SVM) is one of the recent methods for statistical learning, it addresses classification and regression problems . It can be considered as an alternative to neural networks. The advantage of SVM, with respect to neural network, is that it provides a theoretical framework for taking into account not only the experimental data to design an optimal classifier, but also a structural behavior for allowing better generalization capability. This paper introduces SVM theory, applications and its algorithmic implementations. Although there are proven algorithms for constructing SVM programs, it is usually faster and also more reliable to make use or adapt a public domain SVM implementation packages. In this paper, we explain three of the popularly used C/C++ based SVM packages and demonstrate their usage. We report some results of their usage in classification on a number of different datasets, taking into consideration the tuning of SVM kernel hyperparameters for perf...

Series in Machine Perception and Artificial Intelligence, 1999

… Processing, Proceedings of the …, 1994

... section of the bar code, a vision-based reader could reconstruct an average section using the... more ... section of the bar code, a vision-based reader could reconstruct an average section using the whole code area. This is the principle of the proposed approach namely the projection of thebar code's image to a one-dimensional signal followed by the signal processing and its ...

This paper introduces a novel method for online writer identification. Traditional methods make u... more This paper introduces a novel method for online writer identification. Traditional methods make use of the distribution of directions in handwritten traces. The novelty of this paper comes from 1)We propose a text-independent writer identification that uses handwriting stroke's probability distribution function (SPDF) as writer features; 2)We extract four dynamic features to characterize writer individuality; 3)We develop new distance measurement and combine dynamic features in reducing the number of characters required for online text-independent writer identification. In particular, we performed comparative studies of different similarity measures in our experiments. Experiments were conducted on the NLPR handwriting database involving 55 persons. The results show that the new method can improve the identification accuracy and reduce the number of characters required.