Off-Line Handwritten Character Recognition System Using Support Vector Machine

Katiyar, Gauri

doi:10.11648/j.ajnna.20170302.12

Cited by 14 publications

(13 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…After the CNN feature extraction phase, the processed feature sequence is passed to the Recurrent Neural Network (RNN) component. Specifically, we employ a bidirectional Long Short-Term Memory (Bi-LSTM) layer with 256 LSTM units in both forward and backward directions, as shown in Figure 2 (a), totaling 128 units in each direction using equations (1)(2)(3)(4)(5)(6). This bidirectional architecture allows the model to effectively capture dependencies between data entries in the sequence.…”

Section: -Learning Algorithmmentioning

confidence: 99%

“…This processing phase also includes resizing the images to create a standardized dataset, eliminating the requirement for costly recurrent symbol alignment and establishing a consistent baseline for analysis, simplifying the data organization process. [5][6].…”

Section: Introductionmentioning

confidence: 99%

“…A diverse array of features greatly enhances the recognition of handwritten characters. The extraction of robust features is key to representing handwritten text during the feature extraction stage [6]. Training and testing times can be reduced by reducing the number of features and focusing on relevant ones [7].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Handwritten Arabic Bills Reader and Recognizer

Sweidan,

Hammam

2024

Journal of Computing and Communication

View full text Add to dashboard Cite

In pursuit of Egypt's Vision 2030, which emphasizes the pivotal role of governance within state institutions and society, artificial intelligence (AI) stands as a transformative force. A central tenet of this vision involves harnessing AI technologies to accelerate the digitization of documents and their seamless integration into a unified system. This fosters more informed decision-making processes and revolutionizes processing and utilizing information. Our current research project aligns with this broader goal by deep learning capabilities to support organizations involved in enterprise applications. By incorporating AI-driven solutions, we aim to empower these organizations to manage their operations and optimize resource allocation efficiently. The proposed model eliminates manual input of handwritten invoices into ERP applications, resulting in substantive cost savings. Furthermore, the proposed model integrates an entity classification system enhanced by LSTM, significantly improving invoice data's clarity and accuracy. This streamlined approach saves valuable time and enhances the overall effectiveness of resource allocation and decision-making processes. In essence, by integrating AI into document management and enterprise operations, we are not only contributing to the realization of Egypt's vision but also spearheading a technological transformation that has far-reaching implications for governance, efficiency, and progress in the digital age.

show abstract

Section: -Learning Algorithmmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Handwritten Arabic Bills Reader and Recognizer

Sweidan,

Hammam

2024

Journal of Computing and Communication

View full text Add to dashboard Cite

show abstract

“…Pradeep et al [9] applied ANN in handwritten character recognition of 38 English alphabets and numbers. The experiment by Katiyar et al [10] used a well-known standard database acquired from CEDAR that contains images of alphabetic and numeric characters and for evaluating accuracy, SVM was used.…”

Section: Related Workmentioning

confidence: 99%

A New Hybrid Convolutional Neural Network and eXtreme Gradient Boosting Classifier for Recognizing Handwritten Ethiopian Characters

et al. 2020

View full text Add to dashboard Cite

Handwritten character recognition has been profoundly studied for many years in the field of pattern recognition. Due to its vast practical applications and financial implications, the handwritten character recognition is still an important research area. In this research, a Handwritten Ethiopian Character Recognition (HECR) dataset is prepared to train a model. Images in the HECR dataset were organized with more than one color pen RGB main spaces that are size normalized to 28 × 28 pixels. The dataset is a combination of scripts (Fidel in Ethiopia), numerical representations, punctuations, tonal symbols, combining symbols, and special characters. These scripts have been used to write ancient histories, science, and arts of Ethiopia and Eritrea. In this study, a hybrid model of two super classifiers: Convolutional Neural Network (CNN), as well as eXtreme Gradient Boosting (XGBoost), are proposed for classification. In this integrated model, CNN works as a trainable automatic feature extractor from the raw images and XGBoost takes the extracted features as an input for recognition and classification. The output error rates of the hybrid model and CNN with a fully connected layer are compared. A 0.4630 and 0.1612 error rates were achieved in classifying the handwritten testing dataset images, respectively. The XGBoost as a classifier gave better results than the traditional fully connected layer.

show abstract

“…The model was compared with LeNet-5 architecture and obtained a higher accuracy result. 95% accuracy was obtained for this method.SVM based character recognition has been proposed by GauriKatiyar et al [22]. The method emphasizes the advantage of using SVM classifier for character classification.…”

Section: Introductionmentioning

confidence: 99%

Malayalam Handwritten Character Recognition Using AlexNet Based Architecture

James¹,

Manjusha²,

Chandran

2018

IJEEI

View full text Add to dashboard Cite

This research article proposes a new handwritten Malayalam character recognition model based on AlexNet based architecture. The Malayalam language consists of a variety of characters having similar features, thus, differentiating characters is a challenging task. A lot of handcrafted feature extraction methods have been used for the classification of Malayalam characters. Convolutional Neural Networks (CNN) is one of the popular methods used in image and language recognition. AlexNet based CNN is proposed for feature extraction of basic and compound Malayalam characters. Furthermore, Support Vector Machine (SVM) is used for classification of the Malayalam characters. The 44 primary and 36 compound Malayalam characters are recognised with better accuracy and achieved minimal time consumption using this model. A dataset consisting of about 180,000 characters is used for training and testing purposes. This proposed model produces an efficiency of 98% with the dataset. Further, a dataset for Malayalam characters is developed in this research work and shared on Internet.

show abstract

Off-Line Handwritten Character Recognition System Using Support Vector Machine

Cited by 14 publications

References 19 publications

Handwritten Arabic Bills Reader and Recognizer

Handwritten Arabic Bills Reader and Recognizer

A New Hybrid Convolutional Neural Network and eXtreme Gradient Boosting Classifier for Recognizing Handwritten Ethiopian Characters

Malayalam Handwritten Character Recognition Using AlexNet Based Architecture

Contact Info

Product

Resources

About