The A2iA Multi-lingual Text Recognition System at the Second Maurdor Evaluation

Moysset, Bastien; Bluche, Théodore; Knibbe, Maxime; BenZeghiba, Mohamed Faouzi; Messina, Ronaldo; Louradour, Jérôme; Kermorvant, Christopher

doi:10.1109/icfhr.2014.57

Cited by 41 publications

(26 citation statements)

References 16 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Optical Model (OM) was a Multi-directional Long Short-Term Memory (MDLSTM) [9] Neural Network (NN), as described in [10]. The output are the word-tokens recognised by constraining the character predictions from the OM by a Language Model (LM).…”

Section: Competition Protocolmentioning

confidence: 99%

ICFHR2014 Competition on Handwritten Text Recognition on Transcriptorium Datasets (HTRtS)

Sánchez

Romero

Toselli

et al. 2014

2014 14th International Conference on Frontiers in Handwriting Recognition

View full text Add to dashboard Cite

IEEESánchez Peiró, JA.; Romero Gómez, V.; Toselli, AH.; Vidal Ruiz, E. (2014) Abstract-A contest on Handwritten Text Recognition organised in the context of the ICFHR 2014 conference is described. Two tracks with increased freedom on the use of training data were proposed and three research groups participated in these two tracks. The handwritten images for this contest were drawn from an English data set which is currently being considered in the tranScriptorium project. The the goal of this project is to develop innovative, efficient and cost-effective solutions for the transcription of historical handwritten document images, focusing on four languages: English, Spanish, German and Dutch. For the English language, the so-called "Bentham collection" is being considered in tranScriptorium. It encompasses a large set of manuscripts written by the renowned English philosopher and reformer Jeremy Bentham (1748-1832). A small subset of this collection has been chosen for the present HTR competition. The selected subset has been written by several hands (Bentham himself and his secretaries) and entails significant varibilities and difficulties regarding the quality of text images and writting styles. Training and test data were provided in the form of carefully segmented line images, along with the corresponding transcripts. The three participants achieved very good results, with transcription word error rates ranging from 15.0% down to 8.6%.

show abstract

Section: Competition Protocolmentioning

confidence: 99%

ICFHR2014 Competition on Handwritten Text Recognition on Transcriptorium Datasets (HTRtS)

Sánchez

Romero

Toselli

et al. 2014

2014 14th International Conference on Frontiers in Handwriting Recognition

View full text Add to dashboard Cite

show abstract

“…First, in the line of what has been proposed in image classification [6], data augmentation can be performed to enlarge the number of training samples. Real images can be slanted and stretched to form new samples [3] or they can be warped with a random grid-based distortion [7]. Secondly, the training set can be extended by adding to it artificial images.…”

Section: A State Of the Artmentioning

confidence: 99%

Manifold Mixup Improves Text Recognition with CTC Loss

Moysset¹,

Messina²

2019

2019 International Conference on Document Analysis and Recognition (ICDAR)

Self Cite

View full text Add to dashboard Cite

Modern handwritten text recognition techniques employ deep recurrent neural networks. The use of these techniques is especially efficient when a large amount of annotated data is available for parameter estimation. Data augmentation can be used to enhance the performance of the systems when data is scarce. Manifold Mixup is a modern method of data augmentation that meld two images or the feature maps corresponding to these images and the targets are fused accordingly. We propose to apply the Manifold Mixup to text recognition while adapting it to work with a Connectionist Temporal Classification cost. We show that Manifold Mixup improves text recognition results on various languages and datasets.

show abstract

“…Our implementation of this model scans the image in four different directions and produces predictions for each character in the given label alphabet. Details on the architecture and on the training of the model are given in [5].…”

Section: Optical Modelmentioning

confidence: 99%

“…These systems are based on the Recurrent Neural Network (RNN) system used during the Maurdor competition [5]. We present on the one hand a system in which printed and handwritten characters share the same network labels and, on the other hand, a system in which the different write-types have distinct output labels.…”

Section: Introductionmentioning

confidence: 99%

A Comparison of Recognition Strategies for Printed/Handwritten Composite Documents

Moysset¹,

Messina²,

Kermorvant³

2014

2014 14th International Conference on Frontiers in Handwriting Recognition

Self Cite

View full text Add to dashboard Cite

Full-page segmentation and recognition of realworld documents is a challenging task, involving the segmentation of the images (graphics, text) and the subsequent recognition of the detected text-zones. Often those documents present zones with both write-types: printed and handwritten, which so far have been dealt with by classifying the zones according to the write-type and then using type-specific models for recognition.Here we present two recognition systems using state-of-theart recurrent neural networks, that can recognize the text in zones with both write-types, without the need of explicit type identification; just the segmentation in lines is needed. In one of the systems, there is no distinction on the type at the network's output (one output label per character) while in the other there is one output label for each character and write-type.Experiments have been done on real-world documents from the Maurdor competition. These two systems perform at a similar level than systems using specific networks per type on the constrained task where there is only one write-type per zone. They perform better when both handwritten and printed text are present in the text zone.The results open the perspective to treat OCR and handwritten text recognition with a single optical model.

show abstract

The A2iA Multi-lingual Text Recognition System at the Second Maurdor Evaluation

Cited by 41 publications

References 16 publications

ICFHR2014 Competition on Handwritten Text Recognition on Transcriptorium Datasets (HTRtS)

ICFHR2014 Competition on Handwritten Text Recognition on Transcriptorium Datasets (HTRtS)

Manifold Mixup Improves Text Recognition with CTC Loss

A Comparison of Recognition Strategies for Printed/Handwritten Composite Documents

Contact Info

Product

Resources

About