scholarly journals Printed Persian Subword Recognition Using Wavelet Packet Descriptors

2013 ◽  
Vol 2013 ◽  
pp. 1-11 ◽  
Author(s):  
Samira Nasrollahi ◽  
Afshin Ebrahimi
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
Wavelet Packet ◽  
Recognition Rate ◽  
Text Documents ◽  
New Approach ◽  
Feature Vectors ◽  
Invariant Features ◽  
Optical Character ◽  
The Mean

In this paper, we present a new approach to offline OCR (optical character recognition) for printed Persian subwords using wavelet packet transform. The proposed algorithm is used to extract font invariant and size invariant features from 87804 subwords of 4 fonts and 3 sizes. The feature vectors are compressed using PCA. The obtained feature vectors yield a pictorial dictionary for which an entry is the mean of each group that consists of the same subword with 4 fonts in 3 sizes. The sets of these features are congregated by combining them with the dot features for the recognition of printed Persian subwords. To evaluate the feature extraction results, this algorithm was tested on a set of 2000 subwords in printed Persian text documents. An encouraging recognition rate of 97.9% is got at subword level recognition.

scholarly journals Recognition of Handwritten Characters using Deep Convolutional Neural Network

2019 ◽  
Vol 8 (6S4) ◽  
pp. 314-317
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
Recognition Rate ◽  
Writing Style ◽  
Handwritten Character Recognition ◽  
New Approach ◽  
Optical Character ◽  
Handwritten Character ◽  
Handwritten Recognition ◽  
Difficult Part

Handwritten character recognition (HCR) mainly entails optical character recognition. However, HCR involves in formatting and segmentation of the input. HCR is still an active area of research due to the fact that numerous verification in writing style, shape, size to individuals. The main difficult part of Indian handwritten recognition has overlapping between characters. These overlapping shaped characters are difficult to recognize that may lead to low recognition rate. These factors also increase the complexity of handwritten character recognition. This paper proposes a new approach to identify handwritten characters for Telugu language using Deep Learning (DL). The proposed work can be enhance the recognition rate of individual characters. The proposed approach recognizes with overall accuracy is 94%.

scholarly journals Development of the documents comparison module for an electronic document management system

2019 ◽  
pp. 527-533
Author(s):  
M A Mikheev ◽  
P Y Yakimov
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
Document Management ◽  
Electronic Document ◽  
Text Documents ◽  
Text Document ◽  
Document Management System ◽  
Optical Character ◽  
Electronic Document Management ◽  
Scanned Image

The article is devoted to solving the problem of document versions comparison in electronic document management systems. Systems-analogues were considered, the process of comparing text documents was studied. In order to recognize the text on the scanned image, the technology of optical character recognition and its implementation — Tesseract library were chosen. The Myers algorithm is applied to compare received texts. The software implementation of the text document comparison module was implemented using the solutions described above.

Search-Based Classification for Offline Tifinagh Alphabets Recognition

2020 ◽  
pp. 255-267
Author(s):  
Mohammed Erritali ◽  
Youssef Chouni ◽  
Youssef Ouadid
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
Processing Time ◽  
Recognition Rate ◽  
Main Difficulty ◽  
Optical Character

The main difficulty in developing a successful optical character recognition (OCR) system lies in the confusion between the characters. In the case of Amazigh writing (Tifinagh alphabets), some characters have similarities based on rotation or scale. Most of the researchers attempted to solve this problem by combining multiple descriptors and / or classifiers which increased the recognition rate, but at the expense of processing time that becomes more prohibitive. Thus, reducing the confusion of characters and their recognition times is the major challenge of OCR systems. In this chapter, the authors present an off-line OCR system for Tifinagh characters.

scholarly journals APPLICATION OF ZONAL AND CURVATURE FEATURES TO NUMERALS RECOGNITION

2021 ◽  
Vol 9 (2) ◽  
pp. 7-12
Author(s):  
Binod Kumar Prasad
Keyword(s):  
Language Processing ◽  
Character Recognition ◽  
Optical Character Recognition ◽  
Recognition Rate ◽  
Recognition System ◽  
Signature Verification ◽  
Optical Character ◽  
Knn Classifier ◽  
Average Recognition Rate ◽  
Distance Coding

Purpose of the study: The purpose of this work is to present an offline Optical Character Recognition system to recognise handwritten English numerals to help automation of document reading. It helps to avoid tedious and time-consuming manual typing to key in important information in a computer system to preserve it for a longer time. Methodology: This work applies Curvature Features of English numeral images by encoding them in terms of distance and slope. The finer local details of images have been extracted by using Zonal features. The feature vectors obtained from the combination of these features have been fed to the KNN classifier. The whole work has been executed using the MatLab Image Processing toolbox. Main Findings: The system produces an average recognition rate of 96.67% with K=1 whereas, with K=3, the rate increased to 97% with corresponding errors of 3.33% and 3% respectively. Out of all the ten numerals, some numerals like ‘3’ and ‘8’ have shown respectively lower recognition rates. It is because of the similarity between their structures. Applications of this study: The proposed work is related to the recognition of English numerals. The model can be used widely for recognition of any pattern like signature verification, face recognition, character or word recognition in another language under Natural Language Processing, etc. Novelty/Originality of this study: The novelty of the work lies in the process of feature extraction. Curves present in the structure of a numeral sample have been encoded based on distance and slope thereby presenting Distance features and Slope features. Vertical Delta Distance Coding (VDDC) and Horizontal Delta Distance Coding (HDDC) encode a curve from vertical and horizontal directions to reveal concavity and convexity from different angles.

scholarly journals Optical Character Recognition from Printed Text Images

2019 ◽  
pp. 597-604 ◽  
Author(s):  
Dr. T. Kameswara Rao ◽  
K. Yashwanth Chowdary ◽  
I. Koushik Chowdary ◽  
K. Prasanna Kumar ◽  
Ch. Ramesh
Keyword(s):  
Character Recognition ◽  
Optical Character Recognition ◽  
Corner Point ◽  
Document Retrieval ◽  
Document Images ◽  
Image Region ◽  
Text Documents ◽  
Optical Character ◽  
Simple Program ◽  
Point Detection

In recent years, text extraction from document images is one of the most widely studied topics in Image Analysis and Optical Character Recognition. These extractions of document images can be used for document analysis, content analysis, document retrieval and many more. Many complex text extracting processes Maximization Likelihood (ML), Edge point detection, Corner point detection etc. are used to extract text documents from images. In this article, the corner point approach was used. To extract document from images we used a very simple approach based on FAST algorithm. Firstly, we divided the image into blocks and their density in each block was checked. The denser blocks were labeled as text blocks and the less dense were the image region or noise. Then we check the connectivity of the blocks to group the blocks so that the text part can be isolated from the image. This method is very fast and versatile, it can be used to detect various languages, handwriting and even images with a lot of noise and blur. Even though it is a very simple program the precision of this method is closer or higher than 90%. In conclusion, this method helps in more accurate and less complex detection of text from document images.

Application of Mathematical Morphology on Touching or Broken Characters Processing

2010 ◽  
Vol 171-172 ◽  
pp. 73-77
Author(s):  
Ying Jie Liu ◽  
Fu Cheng You
Keyword(s):  
Image Processing ◽  
Mathematical Morphology ◽  
Digital Image Processing ◽  
Character Recognition ◽  
Optical Character Recognition ◽  
Recognition Rate ◽  
Binary Image ◽  
Practical Applications ◽  
Optical Character ◽  
Individual Character

It is difficult to process touching or broken characters in practical applications on optical character recognition. For touching or broken characters, a method based on mathematical morphology of binary image is put forward in the paper. On the basis of the relative theories of digital image processing, the overall process is introduced including separation of touching characters and connection of broken characters. First of all, character image is pre-processed through smoothing and threshold segmentation in order to generate binary image of characters. Then character regions which are touching or broken are processed through different operators of mathematical morphology of binary image by different structuring elements. Thus the touching characters are separated and broken characters are connected. For higher recognition rate, further processes are done to achieve normal and individual character regions.

scholarly journals Demystification of Bilingual Optical Character Recognition System for Devanagari and English Scripts

2019 ◽  
Vol 9 (1) ◽  
pp. 5180-5185
Keyword(s):  
Pattern Recognition ◽  
Character Recognition ◽  
Optical Character Recognition ◽  
Recognition System ◽  
Research Paper ◽  
New Approach ◽  
Optical Character ◽  
New Ideas

Research is deliberately going on in the field of pattern recognition. New ideas are developed and implemented in this field throughout the globe. Optical Character Recognition (OCR) is one of the inseparable applications of Pattern Recognition. Though extensive research is already reported in this field, but multilingual Optical Character Recognition is the most challenging aspect which is still, the need of the hour. Myriads of researchers are digging the information to gather the best solutions for the recognition purpose. In this research paper, we are purposing the steps for the recognition of Devanagari and English scripts simultaneously occurring in the documents. A new approach of segmentation and splitting the characters of both the scripts is also introduced for the benefits of researchers. Most commonly in the documents containing English and Devanagari scripts, English characters are already separated, the challenge is to separate the Devanagari characters. Algorithm to implement the challenging aspect to segment the Devanagari and Roman scripts simultaneously is also implemented in the present paper.

scholarly journals Semantic Text Segmentation from Synthetic Images of Full-Text Documents

2019 ◽  
Vol 18 (6) ◽  
pp. 1381-1406 ◽  
Author(s):  
Lukáš Bureš ◽  
Ivan Gruber ◽  
Petr Neduchal ◽  
Miroslav Hlaváč ◽  
Marek Hrúz
Keyword(s):  
Full Text ◽  
Network Architecture ◽  
Character Recognition ◽  
Optical Character Recognition ◽  
Recognition Rate ◽  
Semantic Segmentation ◽  
Text Documents ◽  
Text Corpora ◽  
Novel Approach ◽  
Synthetic Images

An algorithm (divided into multiple modules) for generating images of full-text documents is presented. These images can be used to train, test, and evaluate models for Optical Character Recognition (OCR). The algorithm is modular, individual parts can be changed and tweaked to generate desired images. A method for obtaining background images of paper from already digitized documents is described. For this, a novel approach based on Variational AutoEncoder (VAE) to train a generative model was used. These backgrounds enable the generation of similar background images as the training ones on the fly.The module for printing the text uses large text corpora, a font, and suitable positional and brightness character noise to obtain believable results (for natural-looking aged documents). A few types of layouts of the page are supported. The system generates a detailed, structured annotation of the synthesized image. Tesseract OCR to compare the real-world images to generated images is used. The recognition rate is very similar, indicating the proper appearance of the synthetic images. Moreover, the errors which were made by the OCR system in both cases are very similar. From the generated images, fully-convolutional encoder-decoder neural network architecture for semantic segmentation of individual characters was trained. With this architecture, the recognition accuracy of 99.28% on a test set of synthetic documents is reached.

Sign in / Sign up

Export Citation Format

Share Document