Paragraph text segmentation into lines with Recurrent Neural Networks

Moysset, Bastien; Kermorvant, Christopher; Wolf, Christian; Louradour, Jérôme

doi:10.1109/icdar.2015.7333803

Cited by 45 publications

(27 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Time taken (sec) Memory (gb) Ours 0.389 ± 0.0888 2.6 Bluche [4] 21.2 ± 5.6 N/a Bluche [5] 0.62 ± 0.14 N/a Wigington [6] 0.546 ± 0.447 9.7 Since our memory usage is substantially smaller, it is possible to run multiple images at the same time; effectively reducing the time required by a third.…”

Section: Methodsmentioning

confidence: 99%

A Computationally Efficient Pipeline Approach to Full Page Offline Handwritten Text Recognition

Chung¹,

Delteil²

2019

2019 International Conference on Document Analysis and Recognition Workshops (ICDARW)

View full text Add to dashboard Cite

Offline handwriting recognition with deep neural networks is usually limited to words or lines due to large computational costs. In this paper, a less computationally expensive full page offline handwritten text recognition framework is introduced. This framework includes a pipeline that locates handwritten text with an object detection neural network and recognises the text within the detected regions using features extracted with a multi-scale convolutional neural network (CNN) fed into a bidirectional long short term memory (LSTM) network. This framework achieves comparable error rates to state of the art frameworks while using less memory and time. The results in this paper demonstrate the potential of this framework and future work can investigate production ready and deployable handwritten text recognisers.

show abstract

Section: Methodsmentioning

confidence: 99%

A Computationally Efficient Pipeline Approach to Full Page Offline Handwritten Text Recognition

Chung¹,

Delteil²

2019

2019 International Conference on Document Analysis and Recognition Workshops (ICDARW)

View full text Add to dashboard Cite

show abstract

“…A huge variety of degradations as well as different resolutions and orientations should be present. Since the landscape of document analysis has changed over the last years, and machine learning based algorithms get more and more popular not only for KWS [19] and HTR [20] but also for LA [21]- [23], the dataset should consist of hundreds of pages to provide an appropriate amount of training samples.…”

Section: Introductionmentioning

confidence: 99%

READ-BAD: A New Dataset and Evaluation Scheme for Baseline Detection in Archival Documents

Grüning

Labahn

Diem

et al. 2018

2018 13th IAPR International Workshop on Document Analysis Systems (DAS)

View full text Add to dashboard Cite

Text line detection is crucial for any application associated with Automatic Text Recognition or Keyword Spotting. Modern algorithms perform good on well-established datasets since they either comprise clean data or simple/homogeneous page layouts. We have collected and annotated 2036 archival document images from different locations and time periods. The dataset contains varying page layouts and degradations that challenge text line segmentation methods. Well established text line segmentation evaluation schemes such as the Detection Rate or Recognition Accuracy demand for binarized data that is annotated on a pixel level. Producing ground truth by these means is laborious and not needed to determine a method's quality. In this paper we propose a new evaluation scheme that is based on baselines. The proposed scheme has no need for binarization and it can handle skewed as well as rotated text lines. The ICDAR 2017 Competition on Baseline Detection and the ICDAR 2017 Competition on Layout Analysis for Challenging Medieval Manuscripts used this evaluation scheme. Finally, we present results achieved by a recently published text line detection algorithm.

show abstract

“…Related to the main line of this PhD Thesis, we found the work of [Moysset et al 2015]. In their approach MDLSTM are trained with Connectionist Temporal Classification (CTC) for the localization of line frontiers.…”

Section: Text Line As Sequencesmentioning

confidence: 99%

“…The image is converted into sequence of feature vectors by dividing the image into D non-overlapping vertical regions, then projection based features are extracted for each block and finally stacked for each row. In this work, they used a HMM to estimate the sequence alignment and the input features of the model are handcrafted by projection profiles, while in [Moysset et al 2015] used the CTC and MDLSTM that are directly applied on the raw image.…”

Section: Text Line As Sequencesmentioning

confidence: 99%

See 1 more Smart Citation

Neural Networks for Document Image and Text Processing

Pellicer¹

View full text Add to dashboard Cite

Paragraph text segmentation into lines with Recurrent Neural Networks

Cited by 45 publications

References 13 publications

A Computationally Efficient Pipeline Approach to Full Page Offline Handwritten Text Recognition

A Computationally Efficient Pipeline Approach to Full Page Offline Handwritten Text Recognition

READ-BAD: A New Dataset and Evaluation Scheme for Baseline Detection in Archival Documents

Neural Networks for Document Image and Text Processing

Contact Info

Product

Resources

About