Character Recognition Task Research Articles

Optical activation function using a metamaterial waveguide for an all-optical neural network.

In this study, we experimentally demonstrated that the nonlinear optical coefficient of the original Si can be enhanced by incorporating a metamaterial structure into an existing silicon waveguide. The two-photon absorption coefficient enhanced by the metamaterial structure was 424 cm/GW, which is 1.2 × 103 times higher than that of Si. Using this metamaterial waveguide-based nonlinear optical activation function, we achieved a high inference accuracy of 98.36% in the handwritten character recognition task, comparable to that obtained with the ReLU function as the activation function. Therefore, our approach can contribute to the realization of more power-efficient and compact all-optical neural networks.

VashaNet: An automated system for recognizing handwritten Bangla basic characters using deep convolutional neural network

Automated character recognition is currently highly popular due to its wide range of applications. Bengali handwritten character recognition (BHCR) is an extremely difficult issue because of the nature of the script. Very few handwritten character recognition (HCR) models are capable of accurately classifying all different sorts of Bangla characters. Recently, image recognition, video analytics, and natural language processing have all found great success using convolutional neural network (CNN) due to its ability to extract and classify features in novel ways. In this paper, we introduce a VashaNet model for recognizing Bangla handwritten basic characters. The suggested VashaNet model employs a 26-layer deep convolutional neural network (DCNN) architecture consisting of nine convolutional layers, six max pooling layers, two dropout layers, five batch normalization layers, one flattening layer, two dense layers, and one output layer. The experiment was performed over 2 datasets consisting of a primary dataset of 5750 images, CMATERdb 3.1.2 for the purpose of training and evaluating the model. The suggested character recognition model worked very well, with test accuracy rates of 94.60% for the primary dataset, 94.43% for CMATERdb 3.1.2 dataset. These remarkable outcomes demonstrate that the proposed VashaNet outperforms other existing methods and offers improved suitability in different character recognition tasks. The proposed approach is a viable candidate for the high efficient practical automatic BHCR system. The proposed approach is a more powerful candidate for the development of an automatic BHCR system for use in practical settings.

Information Extraction from Product Labels: A Machine Vision Approach

This research tackles the challenge of manual data extraction from product labels by employing a blend of computer vision and Natural Language Processing (NLP). We introduce an enhanced model that combines Convolutional Neural Networks (CNN) and Recurrent Neural Networks (RNN) in a Convolutional Recurrent Neural Network (CRNN) for reliable text recognition. Our model is further refined by incorporating the Tesseract OCR engine, enhancing its applicability in Optical Character Recognition (OCR) tasks. The methodology is augmented by NLP techniques and extended through the Open Food Facts API (Application Programming Interface) for database population and text-only label prediction. The CRNN model is trained on encoded labels and evaluated for accuracy on a dedicated test set. Importantly, our approach enables visually impaired individuals to access essential information on product labels, such as directions and ingredients. Overall, the study highlights the efficacy of deep learning and OCR in automating label extraction and recognition.

Augmentation of Additional Arabic Dataset for Jawi Writing and Classification Using Deep Learning

This research aims to create an additional dataset containing Arabic characters for writing Jawi script and to train classification models using deep learning architectures such as InceptionV3 and ResNet34. The initial stage of the study involves digital image processing to obtain the additional Arabic character dataset from several sources, including HMBD, AHAWP, and HUCD, encompassing various connected and disconnected forms of Jawi script. Image processing includes steps such as preprocessing to enhance image quality, segmentation to separate Arabic characters from the background, and augmentation to increase dataset variability. Once the dataset is formed, we train the models using appropriate training data for each InceptionV3 and ResNet34 architecture. The classification evaluation results indicate that the model with ResNet34 architecture achieved the best performance with an accuracy of 96%. This model successfully recognizes Jawi script accurately and consistently, even for classes with similar shapes. The main contribution of this research is the availability of the additional Arabic character dataset that can be utilized for Jawi script recognition and performance assessment of various deep learning models. The study also emphasizes the importance of selecting the appropriate architecture for specific character recognition tasks. The research findings affirm that the model with ResNet34 architecture has excellent capability in recognizing the additional Arabic characters for writing Jawi. The results of this research have the potential to support further developments in Jawi character recognition applications and provide valuable insights for researchers in the field of character recognition sourced from Arabic characters. Dataset augmentation results can be accessed at https://singkat.usk.ac.id/g/En0skCKGAR

Transformative Progress in Document Digitization: An In-Depth Exploration of Machine and Deep Learning Models for Character Recognition

Introduction: this paper explores the effectiveness of character recognition models for document digitization, leveraging diverse machine learning and deep learning techniques. The study, driven by the increasing relevance of image classification in various applications, focuses on evaluating Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), and VGG16 with transfer learning. The research employs a challenging French alphabet dataset, comprising 82 classes, to assess the models' capacity to discern intricate patterns and generalize across diverse characters. Objective: This study investigates the effectiveness of character recognition models for document digitization using diverse machine learning and deep learning techniques. Methods: the methodology initiates with data preparation, involving the creation of a merged dataset from distinct sections, encompassing digits, French special characters, symbols, and the French alphabet. The dataset is subsequently partitioned into training, test, and evaluation sets. Each model undergoes meticulous training and evaluation over a specific number of epochs. The recording of fundamental metrics includes accuracy, precision, recall, and F1-score for CNN, RNN, and VGG16, while SVM and KNN are evaluated based on accuracy, macro avg, and weighted avg. Results: the outcomes highlight distinct strengths and areas for improvement across the evaluated models. SVM demonstrates remarkable accuracy of 98.63%, emphasizing its efficacy in character recognition. KNN exhibits high reliability with an overall accuracy of 97%, while the RNN model faces challenges in training and generalization. The CNN model excels with an accuracy of 97.268%, and VGG16 with transfer learning achieves notable enhancements, reaching accuracy rates of 94.83% on test images and 94.55% on evaluation images. Conclusion: our study evaluates the performance of five models—Support Vector Machine (SVM), K-Nearest Neighbors (KNN), Recurrent Neural Network (RNN), Convolutional Neural Network (CNN), and VGG16 with transfer learning—on character recognition tasks. SVM and KNN demonstrate high accuracy, while RNN faces challenges in training. CNN excels in image classification, and VGG16, with transfer learning, enhances accuracy significantly. This comparative analysis aids in informed model selection for character recognition applications.

You Actually Look Twice At it (YALTAi): using an object detection approach instead of region segmentation within the Kraken engine

Layout Analysis (the identification of zones and their classification) is the first step along line segmentation in Optical Character Recognition and similar tasks. The ability of identifying main body of text from marginal text or running titles makes the difference between extracting the work full text of a digitized book and noisy outputs. We show that most segmenters focus on pixel classification and that polygonization of this output has not been used as a target for the latest competition on historical document (ICDAR 2017 and onwards), despite being the focus in the early 2010s. We propose to shift, for efficiency, the task from a pixel classification-based polygonization to an object detection using isothetic rectangles. We compare the output of Kraken and YOLOv5 in terms of segmentation and show that the later severely outperforms the first on small datasets (1110 samples and below). We release two datasets for training and evaluation on historical documents as well as a new package, YALTAi, which injects YOLOv5 in the segmentation pipeline of Kraken 4.1.

Review on Isolated Urdu Character Recognition: Offline Handwritten Approach

Abstract: This paper summarizes a system for recognizing isolated Urdu characters using advanced machine learning algorithms. The system analyzes visual features of Urdu characters, like strokes and curves, to train models such as CNN, SVM, ANN, and MLP. With a large dataset, the system can accurately predict unseen characters. It can be integrated into various applications for real-time character recognition tasks like OCR (Optical Character Recognition) and handwriting recognition. This literature survey explores research papers focused on character recognition in languages like Urdu, Arabic, Persian, and Sindhi, proposing various techniques like feature extraction, deep learning, and machine learning to enhance character recognition technology. The survey highlights specific studies with high accuracy and discusses recognition systems for Arabic characters as well.

A Novel Siamese Network for Few/Zero-Shot Handwritten Character Recognition Tasks

Deep metric learning is one of the recommended methods for the challenge of supporting few/zero-shot learning by deep networks. It depends on building a Siamese architecture of two homogeneous Convolutional Neural Networks (CNNs) for learning a distance function that can map input data from the input space to the feature space. Instead of determining the class of each sample, the Siamese architecture deals with the existence of a few training samples by deciding if the samples share the same class identity or not. The traditional structure for the Siamese architecture was built by forming two CNNs from scratch with randomly initialized weights and trained by binary cross-entropy loss. Building two CNNs from scratch is a trial and error and time-consuming phase. In addition, training with binary cross-entropy loss sometimes leads to poor margins. In this paper, a novel Siamese network is proposed and applied to few/zero-shot Handwritten Character Recognition (HCR) tasks. The novelties of the proposed network are in. 1) Utilizing transfer learning and using the pre-trained AlexNet as a feature extractor in the Siamese architecture. Fine-tuning a pre-trained network is typically faster and easier than building from scratch. 2) Training the Siamese architecture with contrastive loss instead of the binary cross-entropy. Contrastive loss helps the network to learn a nonlinear mapping function that enables it to map the extracted features in the vector space with an optimal way. The proposed network is evaluated on the challenging Chars74K datasets by conducting two experiments. One is for testing the proposed network in few-shot learning while the other is for testing it in zero-shot learning. The recognition accuracy of the proposed network reaches to 85.6% and 82% in few- and zero-shot learning respectively. In addition, a comparison between the performance of the proposed Siamese network and the traditional Siamese CNNs is conducted. The comparison results show that the proposed network achieves higher recognition results in less time. The proposed network reduces the training time from days to hours in both experiments.

Text extraction and recognition method for license plates

Text extraction from images has always been challenging, especially if the image is taken under bad conditions, like lightning and noise that can influence text detection and recognition. This paper introduces a novel text extraction and recognition technique applied to the case study license plates. The main idea of this study is to detect the license plate in an input image and try to figure out the original country of the car based on the license plate. To accomplish this task, we first started collecting images from the internet, which were about 100 images. Afterward, we extracted the license plate using machine learning methods. Subsequently, we applied k-means clustering as well as thresholding in order to segment the extracted license plate and make the character recognition task easier. Thereafter, a sequence of techniques were applied, such as resizing and cropping the image to limit the wanted area of the desired character we want to extract. The last part of the proposed method is reading the text from the image using EasyOcr method, and using the function find in order to search for the character or the word. his proposed method achieved satisfactory results in detection where we achieved an accuracy of 87%, and a recognition of 97%. As for finding the ‘word’ part, the algorithm succeeded in all the examples.

Structure and Base Analysis of Receptive Field Neural Networks in a Character Recognition Task.

This paper explores extensions and restrictions of shallow convolutional neural networks with fixed kernels trained with a limited number of training samples. We extend the work recently done in research on Receptive Field Neural Networks (RFNN) and show their behaviour using different bases and step-by-step changes within the network architecture. To ensure the reproducibility of the results, we simplified the baseline RFNN architecture to a single-layer CNN network and introduced a deterministic methodology for RFNN training and evaluation. This methodology enabled us to evaluate the significance of changes using the (recently widely used in neural networks) Bayesian comparison. The results indicate that a change in the base may have less of an effect on the results than re-training using another seed. We show that the simplified network with tested bases has similar performance to the chosen baseline RFNN architecture. The data also show the positive impact of energy normalization of used filters, which improves the classification accuracy, even when using randomly initialized filters.

A conditional GAN-based approach for enhancing transfer learning performance in few-shot HCR tasks

Supervised learning with the restriction of a few existing training samples is called Few-Shot Learning. FSL is a subarea that puts deep learning performance in a gap, as building robust deep networks requires big training data. Using transfer learning in FSL tasks is an acceptable way to avoid the challenge of building new deep models from scratch. Transfer learning methodology considers borrowing the architecture and parameters of a previously trained model on a large-scale dataset and fine-tuning it for low-data target tasks. But practically, fine-tuning pretrained models in target FSL tasks suffers from overfitting. The few existing samples are not enough to correctly adjust the pretrained model’s parameters to provide the best fit for the target task. In this study, we consider mitigating the overfitting problem when applying transfer learning in few-shot Handwritten Character Recognition (HCR) tasks. A data augmentation approach based on Conditional Generative Adversarial Networks is introduced. CGAN is a generative model that can create artificial instances that appear more real and indistinguishable from the original samples. CGAN helps generate extra samples that hold the possible variations of human handwriting instead of applying traditional image transformations. These transformations are low-level, data-independent operations, and only produce augmented samples with limited diversity. The introduced approach was evaluated in fine-tuning the three pretrained models: AlexNet, VGG-16, and GoogleNet. The results show that the samples generated by CGAN can enhance transfer learning performance in few-shot HCR tasks. This is by achieving model fine-tuning with fewer epochs and by increasing the model’s F1-score and decreasing the Generalization Error (E_{test}).

A neural network with competitive layers for character recognition

A structure and functioning mechanisms of a neural network with competitive layers are described. The network is intended to solve the character recognition task. The network consists of several competitive layers of neurons. Each layer is a neural network consisting of a number of neurons represented as a layer. The number of neural layers is equal to the number of recognized classes. All neural layers have one-to-one correspondence with one another and with the input raster. The neurons of every layer have mutual lateral learning connections, which weights are modified during the learning process. There is a competitive (inhibitory) relationship between all neural layers. This competitive interaction is realized by means of a “winner-take-all” (WTA) procedure which aim is to select the layer with the highest level of neural activity.Validation of the network has been done in experiments on recognition of handwritten digits of the MNIST database. The experiments have demonstrated that its error rate is few less than 2%, which is not a high result, but it is compensated by rather fast data processing and a very simple structure and functioning mechanisms.

Khmer printed character recognition using attention-based Seq2Seq network

This paper presents an end-to-end deep convolutional recurrent neural network solution for Khmer optical character recognition (OCR) task. The proposed solution uses a sequence-to-sequence (Seq2Seq) architecture with attention mechanism. The encoder extracts visual features from an input text-line image via layers of convolutional blocks and a layer of gated recurrent units (GRU). The features are encoded in a single context vector and a sequence of hidden states which are fed to the decoder for decoding one character at a time until a special end-of-sentence (EOS) token is reached. The attention mechanism allows the decoder network to adaptively select relevant parts of the input image while predicting a target character. The Seq2Seq Khmer OCR network is trained on a large collection of computer-generated text-line images for multiple common Khmer fonts. Complex data augmentation is applied on both train and validation dataset. The proposed model’s performance outperforms the state-of-art Tesseract OCR engine for Khmer language on the validation set of 6400 augmented images by achieving a character error rate (CER) of 0.7% vs 35.9%.

Gradient-Free Neural Network Training via Synaptic-Level Reinforcement Learning

An ongoing challenge in neural information processing is the following question: how do neurons adjust their connectivity to improve network-level task performance over time (i.e., actualize learning)? It is widely believed that there is a consistent, synaptic-level learning mechanism in specific brain regions, such as the basal ganglia, that actualizes learning. However, the exact nature of this mechanism remains unclear. Here, we investigate the use of universal synaptic-level algorithms in training connectionist models. Specifically, we propose an algorithm based on reinforcement learning (RL) to generate and apply a simple biologically-inspired synaptic-level learning policy for neural networks. In this algorithm, the action space for each synapse in the network consists of a small increase, decrease, or null action on the connection strength. To test our algorithm, we applied it to a multilayer perceptron (MLP) neural network model. This algorithm yields a static synaptic learning policy that enables the simultaneous training of over 20,000 parameters (i.e., synapses) and consistent learning convergence when applied to simulated decision boundary matching and optical character recognition tasks. The trained networks yield character-recognition performance comparable to identically shaped networks trained with gradient descent. The approach has two significant advantages in comparison to traditional gradient-descent-based optimization methods. First, the robustness of our novel method and its lack of reliance on gradient computations opens the door to new techniques for training difficult-to-differentiate artificial neural networks, such as spiking neural networks (SNNs) and recurrent neural networks (RNNs). Second, the method’s simplicity provides a unique opportunity for further development of local information-driven multiagent connectionist models for machine intelligence analogous to cellular automata.

Image Classification for Edge-Cloud Setting: A Comparison Study for OCR Application

The increasing number of smart devices has led to a rise in the complexity and volume of the image generated. Deep learning is an increasingly common approach for image classification, a fundamental task in many applications. Due to its high computational requirements, implementation in edge devices becomes challenging. Cloud computing serves as an enabler, allowing devices with limited resources to perform deep learning. For cloud computing, however, latency is an issue and is undesirable. Edge computing addresses the issue by redistributing data and tasks closer to the edge. Still, a suitable offloading strategy is required to ensure optimal performance with methods such as LeNet-5, OAHR, and Autoencoder (ANC) as feature extractors paired with different classifiers (such as artificial neural network (ANN) and support vector machine (SVM)). In this study, models are evaluated using a dataset representing Optical Character Recognition (OCR) task. The OCR application has recently been used in many task-offloading studies. The evaluation is based on the time performance and scoring criteria. In terms of time performance, a fully connected ANN using features from the ANC is faster by a factor of over 60 times compared to the fastest performing SVM. Moreover, scoring performance shows that the SVM is less prone to overfit in the case of a noisy or imbalanced dataset in comparison with ANN. So, adopting SVM in which the data distribution is unspecified will be wiser as there is a lower tendency to overfit. The training and inference time, however, are generally higher than ANN.

BANGLA HANDWRITTEN CHARACTER RECOGNITION USING CONVOLUTION NEURAL NETWORK

Since, last one-decade, numerous deep learning models have been designed to resolve handwritten character recognition task in languages, namely, English, Chinese, Arabic, Japanese and Russian. Recognition of Bengali handwritten character from document image datasets is undoubtedly an open challenging task. Due to the advancement of neural network, many models have been developed and it is improving performance. The LeNet is a pioneering work in the field handwritten document image recognition specially hand written digits from the images by using CNN. This paper focuses on designing a convolution neural network with refinements on layers and its parameter tuning for Bengali character recognition system for classification of 50 different fonts. Our revised CNN model outperforms on some existing approach and shows font-recognition accuracy of 98.46%.

Visual space orientation and representational gravity: Contextual orientation visual cues modulate the perceptual extrapolation of motion.

The perceived offset of a moving target is usually displaced forward, in the direction of motion (representational momentum), and downward, in the direction of gravity (representational gravity). In what refers to the latter, the meaning of "downward in the direction of gravity" is ill-defined, for it is known that the perceived direction of gravity ("downward") results from the interaction of vestibular signals, sensitive to the gravito-inertial vector, an aprioristic tendency to assume that it aligns with the body's main axis (idiotropic vector) and visual cues. The present work aims to disclose what effects visual cues have on representational gravity. Participants performed a spatial localization task as well as a subjective visual vertical (SVV) and an oriented character recognition task (OCHART), with stimuli superimposed on a realistic background either aligned with earth's vertical or tilted rightward or leftward. Outcomes disclosed significant and lawful effects of the orientation of the visual context on spatial localization judgements. Specifically, forward displacement along the target's motion direction was bigger for targets moving along the "horizontal" direction implied by the background scene. These trends were furthermore found to be correlated, at an individual level, with the magnitude of SVV, but not with the perceptual upright (as measured with OCHART). These findings show that features of the spatial localization judgements specifically index the visually induced spatial orientation, thus offering the prospect to expand available tools for inquiries concerning human spatial orientation, besides clarifying the multisensorial nature and significantly expanding the notion of representational gravity. (PsycInfo Database Record (c) 2021 APA, all rights reserved).

MT-YOLOv5: Mobile terminal table detection model based on YOLOv5

Table detection is an important task of optical character recognition(OCR). At present, table detection for desktop applications has basically reached commercial requirements. With the advancement of informatization, personal demand for table detection has gradually increased. There is an urgent need to establish a table detection method that can be deployed on handheld devices. This paper proposes a mobile terminal table detection model based on YOLOv5. First, we used YOLOv5 as the main framework of the model. However, considering the problem of connection redundancy in the backbone of YOLOv5, on the basis of retaining the YOLOv5 multi-scale detection head, we replaced the backbone of YOLOv5 with the same excellent Mobilenetv2. In addition, considering the non-linear defects of the lightweight model, we use deformable convolution to make up for it. This paper has been evaluated on the ICDAR 2019 dataset, and the results show that compared with the baseline model, the model reduces the number of parameters by half and increases the detection speed by 47%. At the same time, the model can reach 35.25 FPS on ordinary Android phones.

Memristive Stochastic Computing for Deep Learning Parameter Optimization

Stochastic Computing (SC) is a computing paradigm that allows for the low-cost and low-power computation of various arithmetic operations using stochastic bit streams and digital logic. In contrast to conventional representation schemes used within the binary domain, the sequence of bit streams in the stochastic domain is inconsequential, and computation is usually non-deterministic. In this brief, we exploit the stochasticity during switching of probabilistic Conductive Bridging RAM (CBRAM) devices to efficiently generate stochastic bit streams in order to perform Deep Learning (DL) parameter optimization, reducing the size of Multiply and Accumulate (MAC) units by 5 orders of magnitude. We demonstrate that in using a 40-nm Complementary Metal Oxide Semiconductor (CMOS) process our scalable architecture occupies 1.55mm$^2$ and consumes approximately 167$\mu$W when optimizing parameters of a Convolutional Neural Network (CNN) while it is being trained for a character recognition task, observing no notable reduction in accuracy post-training.

Direct effects of visual skills and working memory on Chinese character reading in young children

AbstractThe present study aimed to examine how visual skills, verbal working memory, visuospatial working memory, and other general cognitive skills (inhibitory control, attention, and decision speed) were simultaneously correlated with the early acquisition of reading among kindergarten children. A total of 99 Chinese children were tested individually on the Chinese character recognition task, the block design task, the digit span task, the visuospatial sketchpad task, the continuous performance task (CPT), the flanker task, and the choice reaction time task. The results showed that visual skills and working memory accounted for 15% of the unique variance in character reading, whereas other general cognitive skills, including decision speed, inhibitory control and attention, did not significantly explain the variance in character reading when visual skills and working memory were considered. Specifically, verbal working memory, rather than visuospatial working memory, explained the unique variance in children's learning of character reading. These findings highlighted that visual skills and verbal working memory were among the most important precursors of reading in children at the beginning stage of learning Chinese characters.