Accurate, data-efficient, unconstrained text recognition with convolutional neural networks

Yousef, Mohammed; Hussain, Khaled F.; Mohammed, Usama S.

doi:10.1016/j.patcog.2020.107482

Cited by 104 publications

(55 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Our method also outperforms the other ones submitted to the competition on every metric (reported in [8]). While, the method proposed by [18] seems relevant when it is trained without many specific examples, our approach reaches state of the art performance on writer adaptation using 16 specific training pages. Table I also highlights the improvement of each model adaptation (optical and LM).…”

Section: Resultsmentioning

confidence: 88%

“…One of them was using Multi-dimensional LSTM (MDLSTM) [17], which has provided good performance when trained on a generic large data set, but has shown difficulty in carrying the writer adaptation process with few samples. After the competition, [18] proposed a fully convolutional network architecture trained with the CTC loss function for text recognition. This system achieves the best results on the READ 2018 data set, significantly outperforming the other systems.…”

Section: A Optical Model Adaptationmentioning

confidence: 99%

See 1 more Smart Citation

Improving Text Recognition using Optical and Language Model Writer Adaptation

Soullard

Swaileh

Tranouez

et al. 2019

2019 International Conference on Document Analysis and Recognition (ICDAR)

View full text Add to dashboard Cite

State-of-the-art methods for handwriting text recognition are based on deep learning approaches and language modeling that require large data sets during training. In practice, there are some applications where the system processes mono-writer documents, and would thus benefit from being trained on examples from that writer. However, this is not common to have numerous examples coming from just one writer. In this paper, we propose an approach to adapt both the optical model and the language model to a particular writer, from a generic system trained on large data sets with a variety of examples. We show the benefits of the optical and language model writer adaptation. Our approach reaches competitive results on the READ 2018 data set, which is dedicated to model adaptation to particular writers.

show abstract

Section: Resultsmentioning

confidence: 88%

Section: A Optical Model Adaptationmentioning

confidence: 99%

Improving Text Recognition using Optical and Language Model Writer Adaptation

Soullard

Swaileh

Tranouez

et al. 2019

2019 International Conference on Document Analysis and Recognition (ICDAR)

View full text Add to dashboard Cite

show abstract

“…More recently, Yousef et al propose a FCN architecture [8] with an heavy use of normalization through Batch Normalization [21], Batch Renormalization [22] and Layer Normalization [23]. Their system also contains a gated mechanism derived from Highway Networks [24], residual components and Depthwise Separable Convolutions [25].…”

Section: B Recurrence-free Modelsmentioning

confidence: 99%

“…The gating mechanism defined in our proposed architecture is similar to the one presented in [8] but we only used a 2-part split (one accounting for the gate, the other for the features to be selected by the gate), whereas the authors have proposed a 3-part split with a gate composed of two filters bancs with substraction. The split is carried out over the channel axis.…”

Section: Gating Mechanismmentioning

confidence: 99%

“…However, even though they have been steadily enhancing, there is still room for improvement regarding unconstrained offline handwriting recognition. Recently, recurrence-free architectures based on Fully Convolutional Networks with strong improvements such as gated mechanisms, shared-weight layers, residual connections and extensive data augmentation have been proposed [8,9] and seems to achieve state-of-the-art results for handwriting recognition tasks.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Have Convolutions Already Made Recurrence Obsolete for Unconstrained Handwritten Text Recognition?

Coquenet

Soullard

Chatelain

et al. 2019

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

View full text Add to dashboard Cite

Unconstrained handwritten text recognition remains an important challenge for deep neural networks. These last years, recurrent networks and more specifically Long Short-Term Memory networks have achieved state-of-the-art performance in this field. Nevertheless, they are made of a large number of trainable parameters and training recurrent neural networks does not support parallelism. This has a direct influence on the training time of such architectures, with also a direct consequence on the time required to explore various architectures. Recently, recurrence-free architectures such as Fully Convolutional Networks with gated mechanisms have been proposed as one possible alternative achieving competitive results. In this paper, we explore convolutional architectures and compare them to a CNN+BLSTM baseline. We propose an experimental study regarding different architectures on an offline handwriting recognition task using the RIMES dataset, and a modified version of it that consists of augmenting the images with notebook backgrounds that are printed grids.

show abstract