Handwritten Numeral Databases of Indian Scripts and Multistage Recognition of Mixed Numerals
01 Mar 2009-IEEE Transactions on Pattern Analysis and Machine Intelligence (IEEE Computer Society)-Vol. 31, Iss: 3, pp 444-457
TL;DR: P pioneering development of two databases for handwritten numerals of two most popular Indian scripts, a multistage cascaded recognition scheme using wavelet based multiresolution representations and multilayer perceptron classifiers and application for the recognition of mixed handwritten numeral recognition of three Indian scripts Devanagari, Bangla and English.
Abstract: This article primarily concerns the problem of isolated handwritten numeral recognition of major Indian scripts. The principal contributions presented here are (a) pioneering development of two databases for handwritten numerals of two most popular Indian scripts, (b) a multistage cascaded recognition scheme using wavelet based multiresolution representations and multilayer perceptron classifiers and (c) application of (b) for the recognition of mixed handwritten numerals of three Indian scripts Devanagari, Bangla and English. The present databases include respectively 22,556 and 23,392 handwritten isolated numeral samples of Devanagari and Bangla collected from real-life situations and these can be made available free of cost to researchers of other academic Institutions. In the proposed scheme, a numeral is subjected to three multilayer perceptron classifiers corresponding to three coarse-to-fine resolution levels in a cascaded manner. If rejection occurred even at the highest resolution, another multilayer perceptron is used as the final attempt to recognize the input numeral by combining the outputs of three classifiers of the previous stages. This scheme has been extended to the situation when the script of a document is not known a priori or the numerals written on a document belong to different scripts. Handwritten numerals in mixed scripts are frequently found in Indian postal mails and table-form documents.
Citations
More filters
TL;DR: Feature extraction methods are devoid of cumbersome calculations, and classifiers are capable of yielding instantaneous results, therefore, the current system is a real time system.
Abstract: Individuality of handwriting inserts varying curvatures and angles whenever someone writes a sample of a particular numeral which makes the task of its off-line recognition more challenging. The paper addresses both these issues in novel and robust ways by merging two Digital domains, namely Digital Communications and Digital Image Processing. Curvature is treated by finding analytical features based on distance and slope. Distance based treatment is done by means of Delta Distance Coding whereas slope based analysis is executed with Delta Slope Coding. Angular variations have been countered with the help of rotation invariant physical feature i.e., Pixel Moment of Inertia. A due stress has been laid on Pixel Moment of Inertia by finding it both globally and locally in terms of Centroidal Moment of Inertia and Zonal Moment of Inertia respectively. The above mentioned features are further supported with statistical features in order to differentiate very similar looking numeral pairs like (3, 8), (1, 7), (7, 9). Feature extraction methods are devoid of cumbersome calculations, and classifiers are capable of yielding instantaneous results. Therefore, the current system is a real time system. The system has been tested on unconstrained MNIST dataset. The overall recognition accuracy of 99.26% has been obtained.
7 citations
TL;DR: An improvement in recognition result is reported when decision combiner based committee is used along with class related feature selection approach for the recognition of Devanagri handwritten numerals.
Abstract: In this paper novel feature selection approach is used for the recognition of Devanagri handwritten numerals. The numeral images used for the experiments in the study are obtained from standard benchmarking data-set created by CVPR (ISI)Kolkata. The recognition algorithm consists of four basic steps; pre-processing, feature generation, feature subset selection and classification. Features are generated from the boundary of characters, utilizing the direction based histogram of segmented compartment of the character image. The feature selection algorithm is utilizing the concept of information theory and is based on maximum relevance minimum redundancy based objective function. The classification results are obtained for a single neural network based classifier as well as for the committee of Neural Network based classifiers. The paper reports an improvement in recognition result when decision combiner based committee is used along with class related feature selection approach.
7 citations
Cites background from "Handwritten Numeral Databases of In..."
...The benchmarking datasets were not there, but recently some dataset have been developed [12] for isolated numerals....
[...]
...al [12] presented a pioneering effort for the development of handwritten numeral database of Indian scripts....
[...]
...The first benchmark for the offline Hindi Numeral was published in the year 2009 [12]....
[...]
TL;DR: By introducing digits from 10 different languages, MNIST-MIX becomes a more challenging dataset and its imbalanced classification requires a better design of models.
Abstract: In this letter, we contribute a multi-language handwritten digit recognition dataset named MNIST-MIX, which is the largest dataset of the same type in terms of both languages and data samples. With the same data format with MNIST, MNIST-MIX can be seamlessly applied in existing studies for handwritten digit recognition. By introducing digits from 10 different languages, MNIST-MIX becomes a more challenging dataset and its imbalanced classification requires a better design of models. We also present the results of applying a LeNet model which is pre-trained on MNIST as the baseline.
7 citations
Cites background from "Handwritten Numeral Databases of In..."
...The recognition of mixed handwritten numerals of three Indian scripts Devanagari, Bangla and English is considered in [6] and handwritten characters from multilanguage document images, whichmay contain various types of characters, e....
[...]
TL;DR: A solution to the problem of accurate forecasting on the ebb and flow of Vietnam’s Hoabinh Reservoir based on neural network with the Cuckoo Search algorithm is presented and it is expected that this work may be useful for hydrographic forecasting.
Abstract: The accuracy of reservoir flow forecasting has the most significant influence on the assurance of stability and annual operations of hydro-constructions. For instance, accurate forecasting on the ebb and flow of Vietnam’s Hoabinh Reservoir can aid in the preparation and prevention of lowland flooding and drought, as well as regulating electric energy. This raises the need to propose a model that accurately forecasts the incoming flow of the Hoabinh Reservoir. In this study, a solution to this problem based on neural network with the Cuckoo Search (CS) algorithm is presented. In particular, we used hydrographic data and predicted total incoming flows of the Hoabinh Reservoir over a period of 10 days. The Cuckoo Search algorithm was utilized to train the feedforward neural network (FNN) for prediction. The algorithm optimized the weights between layers and biases of the neuron network. Different forecasting models for the three scenarios were developed. The constructed models have shown high forecasting performance based on the performance indices calculated. These results were also compared with those obtained from the neural networks trained by the particle swarm optimization (PSO) and back-propagation (BP), indicating that the proposed approach performed more effectively. Based on the experimental results, the scenario using the rainfall and the flow as input yielded the highest forecasting accuracy when compared with other scenarios. The performance criteria RMSE, MAPE, and R obtained by the CS-FNN in this scenario were calculated as 48.7161, 0.067268 and 0.8965, respectively. These results were highly correlated to actual values. It is expected that this work may be useful for hydrographic forecasting.
7 citations
References
More filters
01 Jan 1998
TL;DR: In this article, a graph transformer network (GTN) is proposed for handwritten character recognition, which can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters.
Abstract: Multilayer neural networks trained with the back-propagation algorithm constitute the best example of a successful gradient based learning technique. Given an appropriate network architecture, gradient-based learning algorithms can be used to synthesize a complex decision surface that can classify high-dimensional patterns, such as handwritten characters, with minimal preprocessing. This paper reviews various methods applied to handwritten character recognition and compares them on a standard handwritten digit recognition task. Convolutional neural networks, which are specifically designed to deal with the variability of 2D shapes, are shown to outperform all other techniques. Real-life document recognition systems are composed of multiple modules including field extraction, segmentation recognition, and language modeling. A new learning paradigm, called graph transformer networks (GTN), allows such multimodule systems to be trained globally using gradient-based methods so as to minimize an overall performance measure. Two systems for online handwriting recognition are described. Experiments demonstrate the advantage of global training, and the flexibility of graph transformer networks. A graph transformer network for reading a bank cheque is also described. It uses convolutional neural network character recognizers combined with global training techniques to provide record accuracy on business and personal cheques. It is deployed commercially and reads several million cheques per day.
42,067 citations
37,017 citations
Book•
16 Jul 1998
TL;DR: Thorough, well-organized, and completely up to date, this book examines all the important aspects of this emerging technology, including the learning process, back-propagation learning, radial-basis function networks, self-organizing systems, modular networks, temporal processing and neurodynamics, and VLSI implementation of neural networks.
Abstract: From the Publisher:
This book represents the most comprehensive treatment available of neural networks from an engineering perspective. Thorough, well-organized, and completely up to date, it examines all the important aspects of this emerging technology, including the learning process, back-propagation learning, radial-basis function networks, self-organizing systems, modular networks, temporal processing and neurodynamics, and VLSI implementation of neural networks. Written in a concise and fluid manner, by a foremost engineering textbook author, to make the material more accessible, this book is ideal for professional engineers and graduate students entering this exciting field. Computer experiments, problems, worked examples, a bibliography, photographs, and illustrations reinforce key concepts.
29,130 citations
TL;DR: In this paper, it is shown that the difference of information between the approximation of a signal at the resolutions 2/sup j+1/ and 2 /sup j/ (where j is an integer) can be extracted by decomposing this signal on a wavelet orthonormal basis of L/sup 2/(R/sup n/), the vector space of measurable, square-integrable n-dimensional functions.
Abstract: Multiresolution representations are effective for analyzing the information content of images. The properties of the operator which approximates a signal at a given resolution were studied. It is shown that the difference of information between the approximation of a signal at the resolutions 2/sup j+1/ and 2/sup j/ (where j is an integer) can be extracted by decomposing this signal on a wavelet orthonormal basis of L/sup 2/(R/sup n/), the vector space of measurable, square-integrable n-dimensional functions. In L/sup 2/(R), a wavelet orthonormal basis is a family of functions which is built by dilating and translating a unique function psi (x). This decomposition defines an orthogonal multiresolution representation called a wavelet representation. It is computed with a pyramidal algorithm based on convolutions with quadrature mirror filters. Wavelet representation lies between the spatial and Fourier domains. For images, the wavelet representation differentiates several spatial orientations. The application of this representation to data compression in image coding, texture discrimination and fractal analysis is discussed. >
20,028 citations
01 Jan 1988
TL;DR: This chapter contains sections titled: The Problem, The Generalized Delta Rule, Simulation Results, Some Further Generalizations, Conclusion.
Abstract: This chapter contains sections titled: The Problem, The Generalized Delta Rule, Simulation Results, Some Further Generalizations, Conclusion
17,604 citations
Additional excerpts
...Ç...
[...]