Emmanuèle Grosicki

ICDAR 2009 Handwriting Recognition Competition

This paper describes the handwriting recognition competitionheld at ICDAR 2009. This competition is based onthe RIMES-database, with French written text documents.These document are classified in three different categories,complete text pages, words, and isolated characters. Thisyear 10 systems were submitted for the handwritten recognitioncompetition on snippets of French words. The systemswere evaluated in three subtask depending of the sizes ofthe used dictionary. A comparison between different classificationand recognition systems show interesting results. Ashort description of the participating groups, their systems,and the results achieved are presented.

Results of the RIMES Evaluation Campaign for Handwritten Mail Processing

This paper presents the results of the second test phase of the RIMES evaluation campaign. The latter is the first large-scale evaluation campaign intended to all the key players of the handwritten recognition and document analysis communities. It proposes various tasks around recognition and indexing of handwritten letters such as those sent by postal mail or fax by individuals to companies or administrations. In this second evaluation test, automatic systems have been evaluated on three themes: layout analysis, handwriting recognition and writer identification. The databases used are part of the RIMES database of 5605 real mails completely annotated as well as secondary databases of isolated characters and handwritten words (250,000 snippets). The paper reports on protocols and gives the results obtained in the campaign.(RIMES : Reconnaissance et Indexation de données Manuscrites et de fac similÉS / Recognition and Indexing of handwritten documents and faxes)

ICDAR 2011 - French Handwriting Recognition Competition

This paper describes the French handwriting recognition competition held at ICDAR 2011. This competition is based on the RIMES-database composed of French written documents corresponding to letters sent by individuals to companies or administrations. Two tasks have been proposed this year : the first one consists in recognizing isolated snippets of words with the help of a given dictionary, the second one consists in recognizing blocks of words segmented into lines. This year 9 systems were submitted for the different competition subtasks. A comparison between different classification and recognition systems show interesting results. A short description of the participating groups, their systems, and the results achieved are presented.

New baseline correction algorithm for text-line recognition with bidirectional recurrent neural networks

Abstract. Many preprocessing techniques have been proposed for isolated word recognition. However, recently, recognition systems have dealt with text blocks and their compound text lines. In this paper, we propose a new preprocessing approach to efficiently correct baseline skew and fluctuations. Our approach is based on a sliding window within which the vertical position of the baseline is estimated. Segmentation of text lines into subparts is, thus, avoided. Experiments conducted on a large publicly available database (Rimes), with a BLSTM (bidirectional long short-term memory) recurrent neural network recognition system, show that our baseline correction approach highly improves performance.

Unconstrained Handwritten Document Layout Extraction Using 2D Conditional Random Fields

The paper describes a new approach using a Conditional Random Fields (CRFs) to extract physical and logical layouts in unconstrained handwritten letters such as those sent by individuals to companies. In this approach, the extraction of the layouts is considered as a labeling task consisting in assigning a label to each pixel of the document image. This label is chosen among a set of labels depicting the layout elements. The CRF-based method models two stochastic processes : the first one corresponds to the association between pixels and labels, the second one to the relationship of one label with respect to its neighboring labels. The CRF model gives access to the global conditional probability of a given labeling of the image according to image features and some prior knowledge about the structure of the document. This global probability is computed by means of local conditional probabilities at each pixel. To find the best label field, a key point of our model is the implementation of the optimal inference 2D Dynamic Programming method. Experiments have been performed on 1250 handwritten letters of the RIMES database. Good results have been reported showing the capacity of our approach to extract simultaneously the physical and logical layouts.

Preliminary experiments in layout analysis of handwritten letters based on textural and spatial information and a 2D Markovian approach

This paper addresses the problem of layout analysis of handwritten letters using textural and spatial information with a bidimensional Markovian approach. In this framework, the layout extraction is viewed as a labeling problem which aims to find the optimal configuration of the Markov random field performed by the 2D dynamic programming method by E. Geoffrois (2003). Preliminary experiments have been led on a small part of the RIMES database by E.Augustine el al. (2006) for which an error rate of 15% has been achieved.

A New Hierarchical Handwritten Document Layout Extraction Based on Conditional Random Field Modeling

In this study we describe a new approach to extract layout of unconstrained handwritten letters such as those sent by individuals to companies. The proposed model uses a hierarchical combination of Conditional Random Fields (CRFs) which gives access to various levels of the layout interpretation. The analysis proceeds by decreasing the resolution and increasing the abstraction of the document, starting from high resolution analysis (pixel level), to a low resolution of the layout structure. Informations of high resolution are used to bring a specific prior knowledge of the layout like presence of textual information. Experiments have been performed on the RIMES database composed of more than 5000 handwritten letters. Good results have been reported showing the capacity of our approach to extract simultaneously the physical and logical layouts.

Comparative Study of HMM and BLSTM Segmentation-Free Approaches for the Recognition of Handwritten Text-Lines

This paper deals with the recognition of free-style handwritten text lines. We compare 2 state-of-the-art segmentation-free recognition approaches. The first one is the popular context-dependent HMM approach (Hidden Markov Models). The second one is the recent BLSTM (Bi-directional Long Short-Term Memory) approach based on recurrent neural networks and memory blocks. For the sake of comparison, both recognizers use the same set of features and language model. They are compared from the following perspectives: sliding window parameters for feature extraction, training and decoding speed and performance accuracy with or without using a language model. We compare these two approaches on the publicly available Rimes database of French handwritten mails. Our main findings are that long frame sequences, obtained with specific window parameters, improve both recognizers, and that BLSTMs outperform HMMs in terms of WER rates, at the expense of considerably longer training times.

Intégration d'informations textuelles de haut niveau en analyse de structures de documents manuscrits non contraints

Cet article decrit une nouvelle approche utilisant des champs aleatoires conditionnels (CAC) pour extraire a la fois la structure physique et la structure logique de documents manuscrits non contraints. De bons resultats ont ete obtenus montrant la capacite des approches CAC a extraire la mise en page d’un document complexe. On se propose dans cet article d’etudier l’apport d’une information textuelle dans la modelisation. On compare cette nouvelle approche avec les approches classiques utilisant uniquement des informations graphiques et spatiales.