A Study of Data Augmentation for Handwritten Character Recognition using Deep Learning

Hayashi, Taihei; Gyohten, Keiji; Ohki, Hidehiro; Takami, Toshihiro

doi:10.1109/icfhr-2018.2018.00102

Cited by 8 publications

(12 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As shown in Table 3, the different recognition levels found in the studies are displayed. In general, the most utilized recognition level was at word level, followed by 5.00 A study of data augmentation for handwritten character recognition using deep learning [55] 5.00 Generative adversarial network based adaptive data augmentation for handwritten Arabic text recognition [56] 5.00 Improving Handwritten Arabic Text Recognition Using an Adaptive Data Augmentation Algorithm [57] 5.00 AFFGANwriting: A Handwriting Image Generation Method Based on Multi-feature Fusion [58] 5.00 text-line level. This indicates that these two levels share similar challenges, where data augmentation applied to words can expand to text lines, and data augmentation applied to text lines can contract to words.…”

Section: Recognition Tasks and Datasetsmentioning

confidence: 99%

“…The CLE database contains 18,000 Urdu ligatures in Unicode format, while the UCOM database comprises 48 distinct lines of Urdu text authored by 100 different writers. For the Japanese language, the simulated "Japanese Handwriting Dataset" (JHD) [32] was adopted in the work of [32], along with the "ETL Character Database" (ETL) [81] in the work of [55]. These last two datasets contain handwritten character images.…”

Section: Recognition Tasks and Datasetsmentioning

confidence: 99%

See 1 more Smart Citation

Data Augmentation for Offline Handwritten Text Recognition: A Systematic Literature Review

de Sousa Neto,

Bezerra,

de Moura

et al. 2024

SN COMPUT. SCI.

View full text Add to dashboard Cite

Offline Handwritten Text Recognition (HTR) systems concern the automatic recognition and transcription of handwritten text from scanned images to digital media. Recently, HTR research field has become increasingly important due to the growing need for digitizing documents and automating data entry across various industries. However, achieving satisfactory results depend on the amount of available samples to train an optical model. Creating and labeling large enough datasets for this purpose often require significant time and effort, that in some situations may be impractical. To address this problem, data augmentation approaches are commonly used as an essential component of HTR systems. In this way, the present work aims to identify, explore, and analyze the scope of data augmentation approaches for offline HTR systems. Furthermore, we detailed our research protocol and answered four pertinent research questions, which enabled us to discuss trends and possible gaps. A search was conducted across five scientific databases, focusing on papers published between 2012 and 2023. The search yielded 976 primary papers, with 32 meeting the criteria for inclusion in this review. Our results indicate that handwriting synthesis is an emerging research field, and we observed that Digital Image Processing (DIP) is still widely used as an image generator. Nevertheless, the application of Generative Adversarial Networks (GAN) has gained traction in recent years owing to its impressive ability to synthesize images of handwritten text with arbitrary style and content. In addition, we explored and analyzed the most commonly used datasets and text recognition levels in the selected works.

show abstract

Section: Recognition Tasks and Datasetsmentioning

confidence: 99%

Section: Recognition Tasks and Datasetsmentioning

confidence: 99%

Data Augmentation for Offline Handwritten Text Recognition: A Systematic Literature Review

de Sousa Neto,

Bezerra,

de Moura

et al. 2024

SN COMPUT. SCI.

View full text Add to dashboard Cite

show abstract

“…Hayashi et al [19] suggested a statistical character structure model for preparing a large amount of image data in the field of HCR. The clue under the proposed model was extracting the strokes that represented the character structure and acquiring their probability distributions.…”

Section: Data Augmentationmentioning

confidence: 99%

A Novel Siamese Network for Few/Zero-Shot Handwritten Character Recognition Tasks

Elaraby¹,

Barakat²,

Rezk³

2023

Computers, Materials &Amp; Continua

View full text Add to dashboard Cite

Deep metric learning is one of the recommended methods for the challenge of supporting few/zero-shot learning by deep networks. It depends on building a Siamese architecture of two homogeneous Convolutional Neural Networks (CNNs) for learning a distance function that can map input data from the input space to the feature space. Instead of determining the class of each sample, the Siamese architecture deals with the existence of a few training samples by deciding if the samples share the same class identity or not. The traditional structure for the Siamese architecture was built by forming two CNNs from scratch with randomly initialized weights and trained by binary cross-entropy loss. Building two CNNs from scratch is a trial and error and time-consuming phase. In addition, training with binary crossentropy loss sometimes leads to poor margins. In this paper, a novel Siamese network is proposed and applied to few/zero-shot Handwritten Character Recognition (HCR) tasks. The novelties of the proposed network are in. 1) Utilizing transfer learning and using the pre-trained AlexNet as a feature extractor in the Siamese architecture. Fine-tuning a pre-trained network is typically faster and easier than building from scratch. 2) Training the Siamese architecture with contrastive loss instead of the binary cross-entropy. Contrastive loss helps the network to learn a nonlinear mapping function that enables it to map the extracted features in the vector space with an optimal way. The proposed network is evaluated on the challenging Chars74K datasets by conducting two experiments. One is for testing the proposed network in few-shot learning while the other is for testing it in zero-shot learning. The recognition accuracy of the proposed network reaches to 85.6% and 82% in few-and zero-shot learning respectively. In addition, a comparison between the performance of the proposed Siamese network and the traditional Siamese CNNs is conducted. The comparison results show that the proposed network achieves higher recognition results in less time. The proposed network reduces the training time from days to hours in both experiments.

show abstract

“…This method calculates probability distribution from the features related to the structure of the character. Then, it generates strokes based on the distribution and forms multitudes of new characters [13].…”

Section: Handwriting Recognition With Data Augmentationmentioning

confidence: 99%

“…Recognizing different types of handwritings requires a large dataset collected from various sources which is both costly and time consuming. One way to deal with this problem is data augmentation [13]. This paper proposes a new data augmentation technique -Rotate, Shift and Stretch (RSS) to generate multitudes of handwriting variations.…”

mentioning

confidence: 99%

An Online Cursive Handwritten Medical Words Recognition System for Busy Doctors in Developing Countries for Ensuring Efficient Healthcare Service Delivery

Tabassum¹,

Rahman²,

Abedin³

et al. 2021

Preprint

View full text Add to dashboard Cite

Doctors in developing countries are too busy to write digital prescriptions. Ninety-seven percent of Bangladeshi doctors write handwritten prescriptions, the majority of which lack legibility. Prescriptions are harder to read as they contain multiple languages. This paper proposes a machine learning approach to recognize doctors' handwriting to create digital prescriptions. A ‘Handwritten Medical Term Corpus’ dataset is developed containing 17,431 samples of 480 medical terms. In order to improve the recognition efficiency, this paper introduces a data augmentation technique to widen the variety and increase the sample size. A sequence of line data is extracted from the augmented images of 1,591,100 samples and fed to a Bidirectional LSTM. Data augmentation includes pattern Rotating, Shifting and Stretching (RSS). Eight different combinations are applied to evaluate the strength of the proposed method. The result shows 93.0% average accuracy (max: 94.5%, min: 92.1%) using Bidirectional LSTM and RSS data augmentation. This accuracy is 19.6% higher than the recognition result with no data expansion. The proposed handwritten recognition technology can be installed in a smartpen for busy doctors which will recognize the writings and digitize them in real-time. It is expected that the smartpen will contribute to reduce medical errors, save medical costs and ensure healthy living in developing countries.

show abstract

A Study of Data Augmentation for Handwritten Character Recognition using Deep Learning

Cited by 8 publications

References 5 publications

Data Augmentation for Offline Handwritten Text Recognition: A Systematic Literature Review

Data Augmentation for Offline Handwritten Text Recognition: A Systematic Literature Review

A Novel Siamese Network for Few/Zero-Shot Handwritten Character Recognition Tasks

An Online Cursive Handwritten Medical Words Recognition System for Busy Doctors in Developing Countries for Ensuring Efficient Healthcare Service Delivery

Contact Info

Product

Resources

About