Structural Off-line Handwriting Character Recognition Using Approximate Subgraph Matching and Levenshtein Distance

Research in offline handwriting recognition for unconstrained text remains a difficult challenge. Some problems such as noise in image, skew of text, cursive letters, and various handwriting style is still an open problem. Many method has been researched to solve those problems, such as k-NN, Neural Network, SVM and HMM. And to improve the recognition result, there are many methods can be implemented in the prerocessing stage. One of them is Kimura's method for slant correction that using chain code for character slant prediction. Those method consumes time and resource in its computation, meanwhile the accuracy is not improved significantly. Therefore, this paper propose to create handwritten character into graph with string representation based on structural approach. The purpose is to provide ability in improving recognition accuracy without relying on normalisation technique. The similarity distance between graphs measured using levenshtein distance. Experiment conducted to recognize handwritten upper-case letters and digits character images which taken from ETL-1 AIST databases. Levenshtein distance has an accuracy of 84.69% on digits and 67.01% on alphabet with 5% size of data for training and value 10 for string representation length. As a comparison, the Kimura's method are implemented for slant correction which results in a reduction of accuracy until 6%. Comparisons also made with some of previous work.

Handwritten character recognition represents a major challenge in the fields of pattern recognition and machine learning, it is more concerned especially in the case of regional languages like Kannada. Kannada, a prominent Dravidian script, poses distinct challenges because of its intricate character formations and diverse handwriting styles. This study introduces a method for identifying handwritten Kannada characters with the application of three machine learning algorithms: K-Nearest Neighbors (KNN), Convolutional Neural Networks (CNN), and Support Vector Machines (SVM). The dataset utilized comprises an extensive array of handwritten Kannada characters sourced from the Kaggle repository. Preprocessing techniques, including image standardization and noise reduction, which are implemented to enhance recognition accuracy. CNN is used for feature extraction as well as classification, while KNN and SVM serve for comparative evaluation. The performance of these models is assessed using metrics such as precision, accuracy, recall, and F1-score. Findings indicate that CNN surpasses KNN and SVM in recognition accuracy, highlights the advantages of deep learning in the classification of handwritten characters. Additionally, a web-based app which is developed using Django to facilitate real-time character recognition. This research significantly contributes to the progress of Optical Character Recognition in Kannada script, promoting advancements in digitization and automation.

The graph has been widely used to model the entities and the relationships among them in real-world applications. Subgraph matching is a core operation in graph data analysis. However, existing exact matching methods may incur high cost as their searched branches are always unpromising. In recent years, several approximate matching solutions have been proposed by exploiting neural networks. Nevertheless, the accuracy of the returned approximate results could be improved significantly. Motivated by these observations, we proposed OptMatch, an efficient and generic neural network-assisted subgraph matching approach, in this work. In particular, OptMatch proposes a novel subgraph partial embedding network and implements carefully designed search strategies to optimize search processes during the subgraph matching process. First, it can be used to accelerate existing exact matching methods. Moreover, it is also an approximate matching solution, which offers better accuracy compared to existing approximate solutions. We conduct exten-sive experiments on seven real-world data graphs to demonstrate the superiority of OptMatch in both exact and approximate subgraph matching.

Offline handwritten character recognition is the process of recognizing given characters from the large set of characters. OCR system mainly focuses on the recognition of printed or handwritten characters of a scanned image. The proposed system extracts features that are based only on gradient of image which is helpful in exact recognition of characters. A technique to recognize handwritten Devanagari characters using combination of quadratic and SVM classifiers is presented in this paper. Features used are directional features that are strength, angle and histogram of gradient (SOG, AOG, HOG). Using a Gaussian filter, the strength and the angle features are down sampled to obtain a feature vector of 392 dimensions. These features are finally concatenated with HOG feature. Applying these to the combination of quadratic and SVM classifiers to obtain maximum accuracy of 95.81% using 3 fold cross validation.

In this paper, we present an embedding-based framework (TrQuery) for recommending solutions of a SPARQL query, including approximate solutions when exact querying solutions are not available due to incompleteness or inconsistencies of real-world RDF data. Within this framework, embedding is applied to score solutions together with edit distance so that we could obtain more fine-grained recommendations than those recommendations via edit distance. For instance, graphs of two querying solutions with a similar structure can be distinguished in our proposed framework while the edit distance depending on structural difference becomes unable. To this end, we propose a novel score model built on vector space generated in embedding system to compute the similarity between an approximate subgraph matching and a whole graph matching. Finally, we evaluate our approach on large RDF datasets DBpedia and YAGO, and experimental results show that TrQuery exhibits an excellent behavior in terms of both effectiveness and efficiency.

A multi-objective approach towards cost effective isolated handwritten Bangla character and digit recognition

Handwritten Chinese character recognition (HCCR) is an important research field of pattern recognition, which has attracted extensive studies during the past decades. Recently convolutional neural network (CNN) based methods have achieved the state-of-the-art performance for handwritten Chinese character recognition. Nevertheless, handwritten Chinese character recognition is still limited to be effectively used in the actual environment due to the large-scale vocabulary and great diversity of handwriting style. In this paper, we constructed a handwritten Chinese character recognition service based on convolutional neural network, which tries to make effective use of handwritten based printed fonts and existing handwritten database. At the same time, the service can effectively collect more handwritten data to expand the training dataset, which makes it easy to adapt to the new handwriting styles. Meanwhile, We propose a multi-level recognition theory applied to online handwritten Chinese character recognition, which may improve the accuracy of handwritten Chinese character recognition and break the limitations of handwritten Chinese character recognition by identifying the structure of Chinese characters and possible stroke orders firstly. Furthermore, we try to apply the method of online character recognition to the offline character recognition based on the basic writing rules.

The present work deals with recognition of handwritten characters of Bangla, a major script of the Indian sub-continent. The main contributions presented here are (a) generation of a database of handwritten basic characters of Bangla and (b) development of a handwritten character recognition scheme suitable for scripts like Bangla consisting of many similar shaped characters for the benchmark results. The present database is a pioneering development in the context of recognition of off-line handwritten characters of this script. It has 37,858 handwritten samples and accommodates a large spectrum of handwriting style by Bangla speaking population. This database will be made available ( http://www.isical.ac.in/~ujjwal/download/Banglabasiccharacter.html ) free of cost to researchers for further studies. Also, we identified two major factors affecting high recognition accuracies for the present character samples, namely, (a) erratic nature of the presence of headline (shapes of Bangla characters usually contain a horizontal line in its upper part) and (b) existence of several pairs of similar shaped characters. The proposed recognition approach takes care of the above factors. It identifies any confusion in the first stage classification between a pair of similar shaped character classes and resolves the same in the second stage classification by extracting a feature vector based on a non-uniform grid.

Efficient methods have been proposed in the literature for recognition of printed characters of Indian scripts. However, much less attention has been received for handwritten characters of Indian scripts with only a few work available in the literature. In this paper, a novel zone based method has been presented for recognition of handwritten characters written in Kannada language, a major South Indian language. The normalized character image is divided into 64 zones each of size 8x8 pixels. For each zone, from left to right and from top to bottom, the crack code, representing the line between the object pixel and the background (the crack), is generated by traversing it in anticlockwise direction. A feature vector of size 512 is obtained for each character. A multi-class SVM is used for the classification purpose. Experiments are performed on handwritten Kannada character images consisting of 24500 images with 500 samples of each character. Five-fold cross validation is used for result computation that yielded 87.24% recognition accuracy.

Handwritten character recognition belongs to the image feature classification problem. To address the problems that convolutional neural network-based handwritten character feature extraction algorithms have complex structures, long training time and recognition accuracy of similar characters is much lower than the average recognition accuracy, we improve the VGG16 and propose a new character recognition algorithm, which effectively improves the recognition accuracy of similar characters. The improved model reduces the complexity and accelerates the model convergence by simplifying the original convolutional structure, adding Ghost convolution, and batch normalization layer. In addition, we introduce an attention mechanism and deformable convolution to solve the problem of false recognition caused by local feature differences and stroke deformation of similar characters. Experimental results on Chinese character datasets and self-built datasets show that our proposed approach achieves over 97% accuracy in general and 92.82% in similar characters. Compared with the traditional approaches, the training speed is faster, and the similar character recognition effect is improved. The model size is reduced by 80%, achieving lightweight and having important application value for improving the performance of character recognition.

Handwritten character recognition has been one of the most fascinating research among the various researches in field of image processing. In Handwritten character recognition method the input is scanned from images, documents and real time devices like tablets, tabloids, digitizers etc. which are then interpreted into digital text. There are basically two approaches — Online Handwritten recognition which takes the input at run time and Offline Handwritten Recognition which works on scanned images. In this paper we have discussed the architecture, the steps involved, and the various proposed methodologies of offline and online character recognition along with their comparison and few applications.

Recent research indicates that recognition of interrupted text can predict speech intelligibility under adverse listening conditions. However, factors underlying the relationship between perceptual processing of speech and text are not fully understood. We examined contributions of underlying linguistic and perceptual structure by comparing recognition of printed and spoken sentences interrupted at different rates (0.5—64 Hz) in 14 normal-hearing adults. The interruption method approximated deletion and retention of rate-specific linguistic information across the two modalities by substituting white space for silent intervals. Results indicate a remarkably similar U-shaped pattern of cross-rate variation for both modalities, with minima at 2 Hz. Nevertheless, at high and low interruption rates text recognition exceeded speech recognition, while the reverse trend was observed at middle rates. Surprisingly, no significant correlations were obtained in recognition accuracy between text and speech conditions. These findings indicate a high degree of perceptual constancy in recognition of interrupted text and speech, which may rely on retention of rate-specific linguistic and perceptual information retained after the interruptions. On the other hand, results also indicate rate-specific variation in perceptual processing of text and speech, which may potentially affect the degree to which recognition accuracy in one modality is predictive of the other.

Most Latin word and character recognition systems have numerous classification algorithms to automate the identification of handwritten characters. However, despite extensive research on various Latin-based languages, relatively limited attention has been given to the recognition of handwritten text in Nigerian indigenous languages. Handwritten character recognition is challenging due to the variations in shape and style. In addition to that, these characters naturally vary among individuals. This work presents an offline character recognition system for Igbo handwritten alphabets using convolutional neural networks. In this study, a dataset comprising 9,000 character images was created, with 250 distinct samples for each of the 36 letters in the Igbo alphabet. Of these, 200 samples per character (a total of 7,200 images) were pre-processed, binarized, and normalized for training two Convolutional Neural Network (CNN) models. The remaining 50 samples per character (1,800 images) were reserved for testing. Model 1 and Model 2 achieved recognition accuracies of 83.5% and 85.0%, respectively. The proposed approach demonstrates potential for enhancing the automation and digitization of handwritten Igbo texts into machine-readable formats. Keywords: Optical Character Recognition, Igbo Handwritten Characters, Nigerian, Indigenous, Convolutional Neural Network Journal Reference Format: Woods, N.C. & Oyedeji, O.A. (2025): Offline Handwritten Character Recognition for Igbo Alphabets Using Convolutional Neural Networks. Journal of Behavioural Informatics, Digital Humanities and Development Rese Vol. 11 No. 2. Pp 115-126 https://www.isteams.net/behavioralinformaticsjournal dx.doi.org/10.22624/AIMS/BHI/V11N2P11

Character Recognition (OCR) for the Madurese language using Genetic Algorithms (GA). The study addresses the challenges in processing Madurese text documents by implementing a nine-step image preprocessing workflow optimized through GA. Our methodology combines rescaling, grayscale conversion, adaptive thresholding, deskewing, median blur, Otsu thresholding, border removal, contrast enhancement, and noise reduction, with the sequence determined by GA optimization. The system utilizes Tesseract 5.5 OCR engine configured with Vietnamese language model parameters to accommodate Maderese writing characteristics. Experiments conducted on a dataset of 500 images demonstrated significant improvements in recognition accuracy. The GA-optimized preprocessing sequence achieved a 24.32% Word Error Rate (WER) and 7.47% Character Error Rate (CER), marking substantial improvements over the baseline Tesseract implementation. Further optimization through language model selection, particularly using the Occitan (OCI) model, yielded 100% accuracy in specific test cases. The research also explored various fitness function configurations, with a 0.7:0.3 WER-to-CER ratio proving most effective. These results demonstrate the potential of GA optimization in enhancing OCR performance for regional languages with unique characteristics, contributing to the broader field of document digitization and language preservation

Chinese characters are constructed by basic strokes based on structural rules. In handwritten characters, the shapes of the strokes may vary to some extent, but the spatial relations and geometric configurations of the strokes are usually maintained. Therefore these spatial relations and configurations could be regarded as invariant features and could be used in the recognition of handwritten Chinese characters. In this paper, we investigate the structural knowledge in Chinese characters and propose the stroke spatial relationship representation (SSRR) to describe Chinese characters. An On-Line Chinese Character Recognition (OLCCR) method using the SSRR is also presented. With SSRR, each character is processed and is represented by an attribute graph. The process of character recognition is thereby transformed into a graph matching problem. After careful analysis, the basic spatial relationship between strokes can be characterized into five classes. A bitwise representation is adopted in the design of the data structure to reduce storage requirements and to speed up character matching. The strategy of hierarchical search in the preclassification improves the recognition speed. Basically, the attribute graph model is a generalized character representation that provides a useful and convenient representation for newly added characters in an OLCCR system with automatic learning capability. The significance of the structural approach of character recognition using spatial relationships is analyzed and is proved by experiments. Realistic testing is provided to show the effectiveness of the proposed method.