Character Recognition Problem Research Articles

Semantic and Context Understanding for Sentiment Analysis in Hindi Handwritten Character Recognition Using a Multiresolution Technique

The rapid growth of Web 2.0, which enables people to generate, communicate, and share information, has resulted in an increase in the total number of users. In developing countries, online users’ sentiment influences decision-making, social views, individual consumption decisions, and entity quality monitoring. As a result, more accurate sentiment analysis, particularly in their native language such as Hindi, is preferred over crude binary categorization. This is because of the abundance of web-based data in Indian languages such as Hindi, Marathi, Kannada, Tamil, and so on. Analyzing this data and recovering valuable and relevant information from handwritten text has become extremely important. Despite years of research and development, no optical writing recognition (OCR) system has ever been certified as completely reliable. The first step in any pattern recognition system is feature selection. In many fields, feature selection is studied as a combinatorial optimization problem. The primary goal of feature selection is to reduce the number of redundant and ineffective traits in the recognition system. This feature selection is used to maintain or improve the performance of the classifier used by the recognition system: A support vector machine (SVM) technique could be used to solve this character recognition problem. The Hindi character recognition system recognizes Hindi characters by employing morphological operations, edge detection, HOG feature extraction, and an SVM-based classifier. The proposed model outperformed the current state-of-the-art method, achieving an accuracy of 96.77%.

Large-scale continual learning for ancient Chinese character recognition

Ancient Chinese character recognition is a challenging problem in the field of pattern recognition. It is difficult to collect all character classes during the training stage due to the numerous classes of ancient Chinese characters and the likelihood of discovering new characters over time. A solution to address this problem is continual learning. However, most continual learning methods are not well-suited for large-scale applications, making them insufficient for solving the problem of ancient Chinese character recognition. Although saving raw data for old classes is a good approach for continual learning to address large-scale problems, it is often infeasible due to the lack of data accessibility in reality. To solve these problems, we propose a large-scale continual learning framework based on the convolutional prototype network (CPN), which does not save raw data for old classes. In this paper, several basic strategies have been proposed for the initial training stage to enhance the feature extraction ability and robustness of the network, which can improve the performance of the model in continual learning. In addition, we propose two practical methods in varying feature space (parameters of feature extractor are changeable) and fixed feature space (parameters of feature extractor are fixed), which enable the model to carry out large-scale continual learning. The proposed method does not save the raw data of old classes and enables simultaneous classification of all existing classes without knowing the incremental batch number. Experiments on the CASIA-AHCDB dataset with 5000 character classes demonstrate the effectiveness and superiority of the proposed method.

Enhanced Recognition of Handwritten Marathi Compound Characters using CNN-SVM Hybrid Approach

This study presents a hybrid recognition system for multi-class compound Marathi characters, which addresses the problem of handwritten Marathi character recognition. The methodology efficiently bridges the gap between feature extraction and classification by integrating a Convolutional Neural Network (CNN) and Support Vector Machine (SVM). The first step is gathering and preprocessing a wide range of handwritten Marathi compound characters that are written in different styles. Using conventional supervised learning methods, the CNN is trained on this dataset, paying special attention to data augmentation and validation in order to reduce overfitting. High-level features taken from the final fully connected layer of the CNN are fed into an SVM classifier in the next step. By using these features in its training, the SVM improves prediction accuracy. For multi-class classification, the one-vs-all method is used. The hybrid CNN-SVM algorithm demonstrates its effectiveness in the crucial phases of feature extraction and classification by identifying handwritten compound Marathi characters with remarkable accuracy. Evaluation metrics, such as accuracy, precision, recall, F1-score, and confusion matrix analysis, are employed in the process of evaluating the effectiveness of the model. This assessment is carried out on a different testing dataset, offering a thorough examination of the model's functionality. The proposed algorithm demonstrates its superior performance and potential for improved character recognition by achieving training accuracy of 98.60% and validation accuracy of 97.69%. The development of handwriting recognition systems has benefited greatly from this research, especially when it comes to intricate scripts like Marathi. The suggested hybrid algorithm shows encouraging outcomes and has a great deal of potential for use in document processing, natural language comprehension, and character recognition in languages that use the Marathi script. Subsequent efforts will centre on refining the model and investigating ensemble methods to increase the robustness and accuracy of recognition.

Reconstructed SqueezeNext with C-CBAM for offline handwritten Chinese character recognition.

Handwritten Chinese character recognition (HCCR) is a difficult problem in character recognition. Chinese characters are diverse and many of them are very similar. The HCCR model consumes a large number of computational resources during runtime, making it difficult to deploy to resource-limited development platforms. In order to reduce the computational consumption and improve the operational efficiency of such models, an improved lightweight HCCR model is proposed in this article. We reconstructed the basic modules of the SqueezeNext network so that the model would be compatible with the introduced attention module and model compression techniques. The proposed Cross-stage Convolutional Block Attention Module (C-CBAM) redeploys the Spatial Attention Module (SAM) and the Channel Attention Module (CAM) according to the feature map characteristics of the deep and shallow layers of the model, targeting enhanced information interaction between the deep and shallow layers. The reformulated intra-stage convolutional kernel importance assessment criterion integrates the normalization nature of the weights and allows for structured pruning in equal proportions for each stage of the model. The quantization aware training is able to map the 32-bit floating-point weights in the pruned model to 8-bit fixed-point weights with minor loss. Pruning with the new convolutional kernel importance evaluation criterion proposed in this article can achieve a pruning rate of 50.79% with little impact on the accuracy rate. The various optimization methods can compress the model to 1.06 MB and achieve an accuracy of 97.36% on the CASIA-HWDB dataset. Compared with the initial model, the volume is reduced by 87.15%, and the accuracy is improved by 1.71%. The model proposed in this article greatly reduces the running time and storage requirements of the model while maintaining accuracy.

Performance Analysis of Embedded Multilayer Perceptron Artificial Neural Networks on Smart Cyber-Physical Systems for IoT Environments.

At present, modern society is experiencing a significant transformation. Thanks to the digitization of society and manufacturing, mainly because of a combination of technologies, such as the Internet of Things, cloud computing, machine learning, smart cyber-physical systems, etc., which are making the smart factory and Industry 4.0 a reality. Currently, most of the intelligence of smart cyber-physical systems is implemented in software. For this reason, in this work, we focused on the artificial intelligence software design of this technology, one of the most complex and critical. This research aimed to study and compare the performance of a multilayer perceptron artificial neural network designed for solving the problem of character recognition in three implementation technologies: personal computers, cloud computing environments, and smart cyber-physical systems. After training and testing the multilayer perceptron, training time and accuracy tests showed each technology has particular characteristics and performance. Nevertheless, the three technologies have a similar performance of 97% accuracy, despite a difference in the training time. The results show that the artificial intelligence embedded in fog technology is a promising alternative for developing smart cyber-physical systems.

Building an Optimal Convolutional Neural Network Model for Solving Complex Character Recognition Problems

The purpose of the study is aimed at developing a lighter architecture of a convolutional neural network model that will cope with the narrowly focused task of recognizing complex characters better than large-scale and well-known ones. As the source data, the characters of the Japanese language are used, consisting of two syllabic alphabets: hiragana and katakana, which are the most complex, since their writing style is characterized by a large number of features and similarity of characters, which greatly complicates the task of their classification and recognition. The author's model of a convolutional neural network is designed in the article, consisting of four convolutional layers, three layers of subdiscretization and three layers of exclusion. The developed model was compared with one of the most popular models of the EfficientNetBO neural network from the point of view of their architecture and the results of work on the same data. To implement its own convolutional neural network model, the classic Keras + Tensorflow bundle was used, since these libraries provide the most convenient tools for working in the field of machine learning. The result of the conducted research is the developed technology of fast and accurate recognition of complex symbols based on a convolutional neural network, which can become the basis of a software product in the field of computer vision.

Comparison of String Similarity Algorithm in post-processing OCR

The Optical Character Recognition (OCR) problem that often occurs is that the image used, has a lot of noise covering letters in a word partially. This can cause misspellings in the process of word recognition or detection in the image. After the OCR process, we must do some post-processing for correcting the word. The words will be corrected using a string similarity algorithm. So what is the best algorithm? We conducted a comparison algorithm including the Levenshtein distance, Hamming distance, Jaro-Winkler, and Sørensen – Dice coefficient. After testing, the most effective algorithm is the Sørensen-Dice coefficient with a value of 0.88 for the value of precision, recall, and F1 score

Research, Direct Construction form Optical Characteristics – OCR using Department Technology

Optical character recognition – using a deep learning technique that describes an overview of optical character recognition, the basic steps in the optical character recognition problem. At the same time, the thesis introduces in detail the Long short term memory (LSTM) deep learning method and its application in optical character recognition. Through testing on a dataset of 10,000 license plate photos, the thesis showed that the application of the LSTM method is quite effective in optical character recognition.

Recognition of compound characters in Kannada language

<p><span lang="EN-US">Recognition of degraded printed compound Kannada characters is a challenging research problem. It has been verified experimentally that noise removal is an essential preprocessing step. Proposed are two methods for degraded Kannada character recognition problem. Method 1 is conventionally used histogram of oriented gradients (HOG) feature extraction for character recognition problem. Extracted features are transformed and reduced using principal component analysis (PCA) and classification performed. Various classifiers are experimented with. Simple compound character classification is satisfactory (more than 98% accuracy) with this method. However, the method does not perform well on other two compound types. Method 2 is deep convolutional neural networks (CNN) model for classification. This outperforms HOG features and classification. The highest classification accuracy is found as 98.8% for simple compound character classification. The performance of deep CNN is far better for other two compound types. Deep CNN turns out to better for pooled character classes.</span></p>

Fast Extreme Learning Machine for Berber Handwritten Latin Script

This article deals with the problem of Berber handwritten character recognition using Extreme Learning Machine. This paradigm has gained significant attention in pattern recognition field thanks to its efficient learning speed and its high accuracy. In this paper, we have used a fast Extreme Learning Machine to recognize efficiently the Latin Berber characters. So, the proposed ELM has been trained over a Berber-MNIST dataset containing images of Amazigh alphabets. This algorithm learns much faster than traditional popular learning algorithms thanks to the use of JAX library which contains several functions to reduce the execution time of our solution. The simulation results show that the handwritten recognition system based on our developed extreme learning machine decreases computational cost and reduces the time required for the whole recognition process. Furthermore, the developed ELM achieves a high performance in terms of recognition accuracy.

Towards a unified framework for identity documents analysis and recognition

Identity documents recognition is far beyond classical optical character recognition problems. Automated ID document recognition systems are tasked not only with the extraction of editable and transferable data but with performing identity validation and preventing fraud, with an increasingly high cost of error. A significant amount of research is directed to the creation of ID analysis systems with a specific focus for a subset of document types, or a particular mode of image acquisition, however, one of the challenges of the modern world is an increasing demand for identity document recognition from a wide variety of image sources, such as scans, photos, or video frames, as well as in a variety of virtually uncontrolled capturing conditions. In this paper, we describe the scope and context of identity document analysis and recognition problem and its challenges; analyze the existing works on implementing ID document recognition systems; and set a task to construct a unified framework for identity document recognition, which would be applicable for different types of image sources and capturing conditions, as well as scalable enough to support large number of identity document types. The aim of the presented framework is to serve as a basis for developing new methods and algorithms for ID document recognition, as well as for far more heavy challenges of identity document forensics, fully automated personal authentication and fraud prevention.

Intelligent system for assessing the harmfulness of food products based on the processing of textual and graphic information

The paper substantiates the need to assess the harm of food for consumers with chronic diseases or allergies, which is important to prevent possible deterioration of the disease or eliminate acute allergic reactions of the human body to harmful ingredients present in the product. It is proved that currently there is no convenient intelligent system that could recognize the composition of products on the Ukrainian market, provide product characteristics and assess the harmfulness of the product. It is proposed to use food labels and packaging as primary sources of food information that is available to the consumer. It is shown that the printed information on the packages is presented in text-graphic form. The development of a mobile system as a software solution for the detection and analysis of textual and graphical information on the composition of products based on the use of artificial intelligence methods is proposed and substantiated. The block diagram of the intelligent mobile system for detection and analysis of food composition has been developed. The MSER algorithm is used to select text regions on the input image matrix in the presented algorithmic software. The solution to the problem of character recognition was based on the use of convolutional neural network MobileNet-V2, which is currently the best option in the classification of images by mobile applications that do not have a server part, and therefore large computing resources. Alignment of text on the image was carried out using the method of finding a rectangle with the smallest area Developed algorithms for grouping words. A decision support algorithm has been proposed to assess the harmfulness of products. The developed system allows personalized selection of food for each individual user, ie, the assessment of the composition of products is calculated taking into account the state of health of use, existing threats, diseases, restrictions or allergies

A Fusion-Based Hybrid-Feature Approach for Recognition of Unconstrained Offline Handwritten Hindi Characters

Hindi is the official language of India and used by a large population for several public services like postal, bank, judiciary, and public surveys. Efficient management of these services needs language-based automation. The proposed model addresses the problem of handwritten Hindi character recognition using a machine learning approach. The pre-trained DCNN models namely; InceptionV3-Net, VGG19-Net, and ResNet50 were used for the extraction of salient features from the characters’ images. A novel approach of fusion is adopted in the proposed work; the DCNN-based features are fused with the handcrafted features received from Bi-orthogonal discrete wavelet transform. The feature size was reduced by the Principal Component Analysis method. The hybrid features were examined with popular classifiers namely; Multi-Layer Perceptron (MLP) and Support Vector Machine (SVM). The recognition cost was reduced by 84.37%. The model achieved significant scores of precision, recall, and F1-measure—98.78%, 98.67%, and 98.69%—with overall recognition accuracy of 98.73%.

Comparative Performance Evaluation of VGG-16 and Capsnet using Kannada MNIST

Abstract: Handwriting recognition is an important problem in character recognition. It is much more difficult especially for regional languages such as Kannada. In this regard there has been a recent surge of interest in designing convolutional neural networks (CNNs) for this problem. However, CNNs typically require large amounts of training data and cannot handle input transformations. Capsule networks, which is referred to as capsNets proposed recently to overcome these shortcomings and posed to revolutionize deep learning solutions. Our particular interest in this work is to recognize kannada digit characters, and making capsnet robust to rotation and transformation. In this paper, we focus to achieve the following objectives :1. Explore whether or not capsnet is capable of providing a better fit for the digit images; 2. Adapt and incorporate capsNets for the problem of kannada MNIST digit classification problem at hand; 3. develop a real time application to take handwritten input from the user and recognize the digit; 4. Compare the capsnet with other models on various parameters. Keywords: Capsule Networks, Deep Learning, Convolutional Neural Networks (CNNs), Kannada MNIST, VGG-16

Handwritten Bangla Numerical Digit Recognition Using Fine Regulated Deep Neural Network

The recognition of handwritten Bangla digit is providing significant progress on optical character recognition (OCR). It is a very critical task due to the similar pattern and alignment of handwriting digits. With the progress of modern research on optical character recognition, it is reducing the complexity of the classification task by several methods, a few problems encounter during recognition and wait to be solved with simpler methods. The modern emerging field of artificial intelligence is the Deep Neural Network, which promises a solid solution to these few handwritten recognition problems. This paper proposed a fine regulated deep neural network (FRDNN) for the handwritten numeric character recognition problem that uses convolutional neural network (CNN) models with regularization parameters which makes the model generalized by preventing the overfitting. This paper applied Traditional Deep Neural Network (TDNN) and Fine regulated deep neural network (FRDNN) models with a similar layer experienced on BanglaLekha-Isolated databases and the classification accuracies for the two models were 96.25% and 96.99%, respectively over 100 epochs. The network performance of the FRDNN model on the BanglaLekha-Isolated digit dataset was more robust and accurate than the TDNN model and depend on experimentation. Our proposed method is obtained a good recognition accuracy compared with other existing available methods.

An Optimized Deep Residual Network with a Depth Concatenated Block for Handwritten Characters Classification

Even though much advancements have been achieved with regards to the recognition of handwritten characters, researchers still face difficulties with the handwritten character recognition problem, especially with the advent of... | Find, read and cite all the research you need on Tech Science Press

A review on existing learning techniques applied in solving optical character recognition problem

Optical character recognition is treated as the classical problem in image processing that can be widely used for application areas related to computer vision. This process is highly dependent on the languages, as well as various formats of information like hand written/printed/sculptured documents as input. The chief operations that can be performed while handling this task are training the system and predicting the characters based on the trained knowledge. Conventionally, classification and regression mechanisms are used for this purpose. Due to the recent advents in technological progress, the same mechanism was served with the help of modern machine learning and deep learning concepts. The main agenda of this paper is to create awareness among the researchers by discussing existing mechanisms related to solve optical character recognition using various machine learning and deep learning concepts.

A Review on Existing Learning Techniques Applied in Solving Optical Character Recognition Problem

Optical character recognition is treated as the classical problem in image processing that can be widely used for application areas related to computer vision. This process is highly dependent on the languages, as well as various formats of information like hand written/printed/sculptured documents as input. The chief operations that can be performed while handling this task are training the system and predicting the characters based on the trained knowledge. Conventionally, classification and regression mechanisms are used for this purpose. Due to the recent advents in technological progress, the same mechanism was served with the help of modern machine learning and deep learning concepts. The main agenda of this paper is to create awareness among the researchers by discussing existing mechanisms related to solve optical character recognition using various machine learning and deep learning concepts.

Oyma Karakterlere Sahip Dairesel Metal Cisimlerin Sınıflandırma Probleminin Görüntü İşleme Yöntemleri İle Çözümü

In this study, two different solution ways have been developed for the problem of classification of industrial small circular metal objects on the surfaces of engraved metal. It is the first proposed solution to perform the pattern matching with XOR operator by extract the character region of the circular metal objects as a pre-process, making the model of the Daugman’s Rubber Sheet Model (DRSM) and performing feature extraction. As a result, obtained that average processing time is 69,72 milliseconds and 0,9398 accuracy rate in the first proposed solution. The second solution is the optical character recognition (OCR) on the circular metal objects that to be realized character region detection and character segmentation as a result of the Maximal Stabil Extremal Region (MSER) and Stroke Width Transform (SWT) algorithms. Character recognition realized by using the model of Convolutional Neural Network (CNN) class which is a deep machine learning approach of artificial intelligence. The character recognition problem of the circular metal objects provided at the same time solved the problem of object classification. As a result, obtained that average processing time is 1,596 second and 0,9719 accuracy rate in the second proposed solution.

Research on fuzzy English automatic recognition and human-computer interaction based on machine learning

Fuzzy English recognition is affected by many factors, which leads to certain accuracy problems in intelligent recognition results. In order to improve the automatic recognition efficiency of fuzzy English, based on machine learning technology, this study constructs a neural network model. At the same time, this paper analyzes the research status and existing problems of handwritten character recognition, analyzes the model, and adopts multiple modules for automatic English recognition. In addition, the system is built on the basis of algorithms and model support, which makes fuzzy English recognition intelligent. Finally, in order to study the algorithm and model performance, the fuzzy English recognition is carried out through experiments. The research shows that the model constructed in this paper has certain recognition effect, which can be applied to practice, and can provide theoretical reference for subsequent related research.