SPAN: A Simple Predict &amp; Align Network for Handwritten Paragraph Recognition

Coquenet, Denis; Chatelain, Clément; Paquet, Thierry

doi:10.1007/978-3-030-86334-0_5

Cited by 22 publications

(19 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…They refer to deep architectures composed of many convolutional layers and no recurrent layer. They often include the most recent innovations in deep learning, like gating mechanism or residual connections to obtain state of the art results [5,4,9,19]. These architectures benefit from the computational parallelism offered by convolutional layers, and hence can be trained faster than traditional architectures based on recurrent layers.…”

Section: Standard Approaches For Htrmentioning

confidence: 99%

A Light Transformer-Based Architecture for Handwritten Text Recognition

Barrere

Soullard

Lemaitre

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Transformer models have been showing ground-breaking results in the domain of natural language processing. More recently, they started to gain interest in many others fields as in computer vision. Traditional Transformer models typically require a significant amount of training data to achieve satisfactory results. However, in the domain of handwritten text recognition, annotated data acquisition remains costly resulting in small datasets compared to those commonly used to train a Transformer-based model. Hence, training Transformer models able to transcribe handwritten text from images remains challenging. We propose a light encoder-decoder Transformer-based architecture for handwriting text recognition, containing a small number of parameters compared to traditional Transformer architectures. We trained our architecture using a hybrid loss, combining the well-known connectionist temporal classification with the cross-entropy. Experiments are conducted on the well-known IAM dataset with and without the use of additional synthetic data. We show that our network reaches state-of-the-art results in both cases, compared with other larger Transformer-based models.

show abstract

Section: Standard Approaches For Htrmentioning

confidence: 99%

A Light Transformer-Based Architecture for Handwritten Text Recognition

Barrere

Soullard

Lemaitre

et al. 2022

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…To alleviate this issue, architectures handling singlecolumn pages (or paragraphs) were recently proposed. In [2], [3], the aim is to reformulate the two-dimensional problem to a one-dimensional problem in order to use the CTC loss, which is only designed for one-dimensional sequence alignment problems. Indeed, this loss enables oneshot predictions, contrary to the cross entropy loss, which implies a recurrent process.…”

Section: Handwritten Text Recognitionmentioning

confidence: 99%

“…This way, prediction times are shorter. While the model from [2] learns a representation transformation, the model from [3] learns to align the predictions on the vertical axis, thanks to a reshaping operation. [1] and [4] proposed attention-based models.…”

Section: Handwritten Text Recognitionmentioning

confidence: 99%

“…The emergence of end-to-end models enabled to alleviate the need for segmentation labels. However, the majority of Manuscript received ...; revised ... the recent state-of-the-art models [1], [2], [3], [4] are limited to the recognition of single columns of text. It means that they can only be used for simple layouts or used as a second step, after a prior paragraph segmentation step.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition

Coquenet¹,

Chatelain²,

Paquet³

2022

Preprint

Self Cite

View full text Add to dashboard Cite

Unconstrained handwritten document recognition is a challenging computer vision task. It is traditionally handled by a two-step approach combining line segmentation followed by text line recognition. For the first time, we propose an end-to-end segmentation-free architecture for the task of handwritten document recognition: the Document Attention Network. In addition to the text recognition, the model is trained to label text parts using begin and end tags in an XML-like fashion. This model is made up of an FCN encoder for feature extraction and a stack of transformer decoder layers for a recurrent token-by-token prediction process. It takes whole text documents as input and sequentially outputs characters, as well as logical layout tokens. Contrary to the existing segmentation-based approaches, the model is trained without using any segmentation label. We achieve competitive results on the READ dataset at page level, as well as double-page level with a CER of 3.53% and 3.69%, respectively. We also provide results for the RIMES dataset at page level, reaching 4.54% of CER. We provide all source code and pre-trained model weights at https://github.com/FactoDeepLearning/DAN.

show abstract

“…1 Then, each line is transcribed independently with line-level HTR. Furthermore, recent works do not explicitly include any LA step and aim to obtain the transcription hypothesis by processing whole pages or paragraphs [4,13,50,40].…”

Section: Introductionmentioning

confidence: 99%

End-to-End Page-Level Assessment of Handwritten Text Recognition

Vidal¹,

Toselli²,

Ríos-Vila³

et al. 2023

Preprint

View full text Add to dashboard Cite

The evaluation of Handwritten Text Recognition (HTR) systems has traditionally used metrics based on the edit distance between HTR and ground truth (GT) transcripts, at both the character and word levels. This is very adequate when the experimental protocol assumes that both GT and HTR text lines are the same, which allows edit distances to be independently computed to each given line. Driven by recent advances in pattern recognition, HTR systems increasingly face the end-to-end page-level transcription of a document, where the precision of locating the different text lines and their corresponding reading order (RO) play a key role. In such a case, the standard metrics do not take into account the inconsistencies that might appear. In this paper, the problem of evaluating HTR systems at the page level is introduced in detail. We analyze the convenience of using a two-fold evaluation, where the transcription accuracy and the RO goodness are considered separately. Different alternatives are proposed, analyzed and empirically compared both through partially simulated and through real, full end-to-end experiments. Results support the validity of the proposed two-fold evaluation approach. An important conclusion is that such an evaluation can be adequately achieved by just two simple and well-known metrics: the Word Error Rate, that takes transcription sequentiality into account, and the here re-formulated Bag of Words Word Error Rate, that ignores order. While the latter directly and very accurately assess intrinsic word recognition errors, the difference between both metrics gracefully correlates with the Spearman's Foot Rule Distance, a metric which explicitly measures RO errors associated with layout analysis flaws.

show abstract

SPAN: A Simple Predict & Align Network for Handwritten Paragraph Recognition

Cited by 22 publications

References 20 publications

A Light Transformer-Based Architecture for Handwritten Text Recognition

A Light Transformer-Based Architecture for Handwritten Text Recognition

DAN: a Segmentation-free Document Attention Network for Handwritten Document Recognition

End-to-End Page-Level Assessment of Handwritten Text Recognition

Contact Info

Product

Resources

About