An Investigative Analysis of Different LSTM Libraries for Supervised and Unsupervised Architectures of OCR Training

Bukhari, Syed Saqib; Francis, Sumam; Kamath, Cannannore Nidhi Narayana; Dengel, Andreas

doi:10.1109/icfhr-2018.2018.00084

Cited by 2 publications

(2 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The RNN is a kind of neural network that is part of deep learning, it allows input data to flow in any direction along a time sequence through the layers that make up the network. The Long short-term memory (LSTM) is a type of RNN that has feedback connections, managing to process not only individual data but also complete sequences of data, avoiding loss of information over time [22]. For the proposed system, we worked with the Tesseract software, which is open source, which uses OCR and is complemented by LSTM to achieve a more efficient result in the field of text recognition through digital images.…”

Section: B) Recurrent Neural Network (Rnn) and Tesseract Ocrmentioning

confidence: 99%

Implementation of a Web System: Prevent Fraud Cases in Electronic Transactions

Ponce¹,

Sanchez²

2022

IJACSA

View full text Add to dashboard Cite

The purpose of this project is to prevent cases of fraud in e-commerce of purchase and sale from person to person through social networks. For the development of the research work, the Scrum methodology was used to allow the project to be carried out in an agile and flexible way, adapting to the changes that could arise along the way. The technological tools that made this project possible were SQL Server, C++, Visual Studio and Marvel app, the latter for prototype design. In addition, there was the support of an artificial intelligence software known as Optical Character Recognition that allowed the document recognition process to be completed. The social network Facebook was also relevant for the development process since the data set for the training of the system was obtained from there, guaranteeing its functionality. The results obtained benefit both parties, sellers/suppliers and consumers, reducing the impact of fraud cases and guaranteeing safer online operations. In addition, a validation was carried out by experts in the development of web applications, taking usability, feasibility, scalability, innovation, and technology as criteria. Obtaining as a result the approval in all its criteria; with the total mean value of 2.76.

show abstract

Section: B) Recurrent Neural Network (Rnn) and Tesseract Ocrmentioning

confidence: 99%

Implementation of a Web System: Prevent Fraud Cases in Electronic Transactions

Ponce¹,

Sanchez²

2022

IJACSA

View full text Add to dashboard Cite

show abstract

“…For the past years, there have been significant developments in the field of optical character recognition (OCR) [1]- [3]. It is one of the vital branches of computer vision research.…”

Section: Introductionmentioning

confidence: 99%

InMAS: Deep Learning for Designing Intelligent Making System

Shao

Yuan

et al. 2019

IEEE Access

View full text Add to dashboard Cite

Deep learning is one of the notable solutions when developing intelligent making systems (InMASs) for students' test papers and assignments to replace the workload of the teachers and educators. This paper recommends a design method of InMAS based on the You Only Look Once (YOLOv3) algorithm. Such a method can be used in carrying out experiments on algorithm problems and creating two dedicated datasets. The first is for localization and the second is for recognition. The YOLOv3 network is employed to identify the location and extraction of each mathematical problem in every image. In the recognition part, because of the low recognition rate of traditional optical character recognition (OCR) on the handwritten characters, the YOLOv3 network is widely used to identify the characters in each arithmetic problem. In the final step, the numerical operation is held on the output characters. The template matching approach is used to evaluate the arithmetic problems with the wrong operation in the original pictures. The experimental results show that the accuracy of localization is near to 1. The proposed method carries out favorably against Baidu OCR based on recognition accuracy, showing a high accuracy of 97.15%.

show abstract

An Investigative Analysis of Different LSTM Libraries for Supervised and Unsupervised Architectures of OCR Training

Cited by 2 publications

References 14 publications

Implementation of a Web System: Prevent Fraud Cases in Electronic Transactions

Implementation of a Web System: Prevent Fraud Cases in Electronic Transactions

InMAS: Deep Learning for Designing Intelligent Making System

Contact Info

Product

Resources

About