scispace - formally typeset
Search or ask a question

Showing papers on "Intelligent word recognition published in 1994"


Journal ArticleDOI
TL;DR: A complete scheme for totally unconstrained handwritten word recognition based on a single contextual hidden Markov model type stochastic network is presented, which includes a morphology and heuristics based segmentation algorithm, a training algorithm that can adapt itself with the changing dictionary.
Abstract: Because of large variations involved in handwritten words, the recognition problem is very difficult. Hidden Markov models (HMM) have been widely and successfully used in speech processing and recognition. Recently HMM has also been used with some success in recognizing handwritten words with presegmented letters. In this paper, a complete scheme for totally unconstrained handwritten word recognition based on a single contextual hidden Markov model type stochastic network is presented. Our scheme includes a morphology and heuristics based segmentation algorithm, a training algorithm that can adapt itself with the changing dictionary, and a modified Viterbi algorithm which searches for the (l+1)th globally best path based on the previous l best paths. Detailed experiments are carried out and successful recognition results are reported. >

258 citations


Journal ArticleDOI
TL;DR: In this paper, the authors described techniques to separate a line of unconstrained (written in a natural manner) handwritten text into words, using original algorithms to determine distances between components in a text line and to detect punctuation.

142 citations


Journal ArticleDOI
TL;DR: An automatic off-line character recognition system for handwritten cursive Arabic characters is presented and proved to be powerful in tolerance to variable writing, speed, and recognition rate.
Abstract: An automatic off-line character recognition system for handwritten cursive Arabic characters is presented. A robust noise-independent algorithm is developed that yields skeletons that reflect the structural relationships of the character components. The character skeleton is converted to a tree structure suitable for recognition. A set of fuzzy constrained character graph models (FCCGM's), which tolerate large variability in writing, is designed. These models are graphs, with fuzzily labeled arcs used as prototypes for the characters. A set of rules is applied in sequence to match a character tree to an FCCGM. Arabic handwritings of four writers were used in the learning and testing stages. The system proved to be powerful in tolerance to variable writing, speed, and recognition rate. >

121 citations


Proceedings ArticleDOI
01 Aug 1994
TL;DR: This study evaluates retrieval effectiveness from OCR text databases using a probabilistic IR system and shows there is no statistical difference in precision and recall using graded accuracy levels from three OCR devices.
Abstract: Character accuracy of optically recognized text is considered a basic measure for evaluating OCR devices. In the broader sense, another fundamental measure of an OCR’s goodness is whether its generated text is usable for retrieving information. In this study, we evaluate retrieval effectiveness from OCR text databases using a probabilistic IR system. We compare these retrieval results to their manually corrected equivalent. We show there is no statistical difference in precision and recall using graded accuracy levels from three OCR devices. However, characteristics of the OCR data have side effects that could cause unstable results with this IR model. In particular, we found individual queries can be greatly affected. Knowing the qualities of OCR text, we compensate for them by applying an automatic post-processing system that improves effectiveness.

108 citations


Proceedings Article
01 Jan 1994
TL;DR: A new modular classification system based on several autoassociative multilayer perceptrons which allows the efficient incorporation of high-level knowledge about the learning problem and compared to other approaches to the invariance problem.
Abstract: When training neural networks by the classical backpropagation algorithm the whole problem to learn must be expressed by a set of inputs and desired outputs. However, we often have high-level knowledge about the learning problem. In optical character recognition (OCR), for instance, we know that the classification should be invariant under a set of transformations like rotation or translation. We propose a new modular classification system based on several autoassociative multilayer perceptrons which allows the efficient incorporation of such knowledge. Results are reported on the NIST database of upper case handwritten letters and compared to other approaches to the invariance problem.

69 citations


BookDOI
01 Jan 1994
TL;DR: This paper presents a meta-modelling approach for practical character recognition system development using a pen-based music editor and a model-based dynamic signature verification system.
Abstract: 1: Introduction and overview of field.- Frontiers in handwriting recognition.- 2: Handwritten character recognition.- Historical review of theory and practice of handwritten character recognition.- Automatic recognition of handwritten characters.- Learning, representation, understanding and recognition of characters and words - an intelligent approach.- Digital transforms in handwriting recognition.- Pattern recognition with optimal margin classifiers.- 3: Handwritten word recognition.- On the robustness of recognition of degraded line images.- Invariant handwriting features useful in cursive script recognition.- Off-line recognition of bad quality handwritten words using prototypes.- Handwriting recognition by statistical methods.- Towards a visual recognition of cursive script.- A hierarchical handwritten word segmentation.- 4: Contextual methods in handwriting recognition.- Cursive words recognition: methods and strategies.- Hidden Markov models in handwriting recognition.- Language-level syntactic and semantic constraints applied to visual word recognition.- Verification of handwritten British postcodes using address features.- Improvement of OCR by language model.- An approximate string matching method for handwriting recognition post-processing using a dictionary.- 5: Neural networks in handwriting recognition.- Neural-net computing for machine recognition of handwritten English language text.- Cooperation of feedforward neural networks for handwritten digit recognition.- Normalisation and preprocessing for a recurrent network off-line handwriting recognition system.- 6: Architectures for handwriting.- Architectures for handwriting recognition.- 7: Databases for handwriting recognition.- Large database organization for document images.- 8: Signature recognition and verification.- A model-based dynamic signature verification system.- Algorithms for signature verification.- Handwritten signature verification: a global approach.- 9: Application of handwriting recognition.- Total approach for practical character recognition system development.- A pen-based music editor.

53 citations



Proceedings ArticleDOI
09 Oct 1994
TL;DR: A complete OCR system is described for documents of single Bangla (Bengali) font by a combination of template and feature matching approach and has an accuracy of about 96%.
Abstract: In this paper a complete OCR system is described for documents of single Bangla (Bengali) font. The character shapes are recognized by a combination of template and feature matching approach. Images digitized by flatbed scanner are subjected to skew correction, line, word and character segmentation, simple and compound character separation, feature extraction and finally character recognition. A feature based tree classifier is used for simple character recognition. Preprocessing like thinning and skeletonization is not necessary in our scheme and hence the system is quite fast. At present, the system has an accuracy of about 96%. Also, some character occurrence statistics have been computed to model an error detection and correction technique in the near future.

49 citations


Proceedings ArticleDOI
23 Mar 1994
TL;DR: A new expert system for automatically correcting errors made by optical character recognition (OCR) devices is described, designed to improve the quality of text produced by an OCR device in preparation for subsequent retrieval from an information system.
Abstract: This paper describes a new expert system for automatically correcting errors made by optical character recognition (OCR) devices. The system, which we call the post-processing system, is designed to improve the quality of text produced by an OCR device in preparation for subsequent retrieval from an information system. The system is composed of numerous parts: an information retrieval system, an English dictionary, a domain-specific dictionary, and a collection of algorithms and heuristics designed to correct as many OCR errors as possible. For the remaining errors that cannot be corrected, the system passes them on to a user-level editing program. This post-processing system can be viewed as part of a larger system that would streamline the steps of taking a document from its hard copy form to its usable electronic form, or it can be considered a stand alone system for OCR error correction. An earlier version of this system has been used to process approximately 10,000 pages of OCR generated text. Among the OCR errors discovered by this version, about 87% were corrected. We implement numerous new parts of the system, test this new version, and present the results.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

44 citations


Patent
11 Jul 1994
TL;DR: In this article, a post-processing method for an optical character recognition (OCR) method for combining different OCR engines to identify and resolve characters and attributes of the characters that are erroneously recognized by multiple OCR recognition engines.
Abstract: A post-processing method for an optical character recognition (OCR) method for combining different OCR engines to identify and resolve characters and attributes of the characters that are erroneously recognized by multiple optical character recognition engines. The characters can originate from many different types of character environments. OCR engine outputs are synchronized in order to detect matches and mismatches between said OCR engine outputs by using synchronization heuristics. The mismatches are resolved using resolution heuristics and neural networks. The resolution heuristics and neural networks are based on observing many different conventional OCR engines in different character environments to find what specific OCR engine correctly identifies a certain character having particular attributes. The results are encoded into the resolution heuristics and neural networks to create an optimal OCR post-processing solution.

42 citations


Proceedings ArticleDOI
19 Apr 1994
TL;DR: The MS-TDNN integrates the high accuracy single character recognition capabilities of a TDNN with a non-linear time alignment procedure (dynamic time warping algorithm) for finding stroke and character boundaries in isolated, handwritten characters and words.
Abstract: Shows how the multi-state time delay neural network (MS-TDNN), which is already used successfully in continuous speech recognition tasks, can be applied both to online single character and cursive (continuous) handwriting recognition. The MS-TDNN integrates the high accuracy single character recognition capabilities of a TDNN with a non-linear time alignment procedure (dynamic time warping algorithm) for finding stroke and character boundaries in isolated, handwritten characters and words. In this approach each character is modelled by up to 3 different states and words are represented as a sequence of these characters. The authors describe the basic MS-TDNN architecture and the input features used in the paper, and present results (up to 97.7% word recognition rate) both on writer dependent/independent, single character recognition tasks and writer dependent, cursive handwriting tasks with varying vocabulary sizes up to 20000 words. >

Proceedings ArticleDOI
09 Oct 1994
TL;DR: The paper proposes a structural technique for automatic recognition of hand printed Arabic characters that is more efficient for large and complex sets such as Arabic characters; not expensive for feature extraction; and its execution time does not depend on either the font or the size of the characters.
Abstract: The paper proposes a structural technique for automatic recognition of hand printed Arabic characters. The advantages of this technique are: more efficient for large and complex sets such as Arabic characters; not expensive for feature extraction; and its execution time does not depend on either the font or the size of the characters. The algorithm was implemented on a microcomputer and tested by 10 different users. The recognition rate obtained was about 90%.

Proceedings Article
01 Jan 1994
TL;DR: A geometrical model of the word spatial structure is fitted to the pen trajectory using the EM algorithm and the fitting process maximizes the likelihood of the trajectory given the model and a set a priors on its parameters.
Abstract: We introduce a new approach to normalizing words written with an electronic stylus that applies to all styles of handwriting (upper case, lower case, printed, cursive, or mixed). A geometrical model of the word spatial structure is fitted to the pen trajectory using the EM algorithm. The fitting process maximizes the likelihood of the trajectory given the model and a set a priors on its parameters. The method was evaluated and integrated to a recognition system that combines neural networks and hidden Markov models.

Proceedings ArticleDOI
09 Oct 1994
TL;DR: A new system for the recognition of handwritten text that goes from raw, binary scanned images of census forms to ASCII transcriptions of the fields contained within the forms, using two different statistical language models.
Abstract: A new system for the recognition of handwritten text is described. The system goes from raw, binary scanned images of census forms to ASCII transcriptions of the fields contained within the forms. The first step is to locate and extract the handwritten input from the forms. Then, a large number of character subimages are extracted and individual classified using a MLP (multilayer perceptron). A Viterbi-like algorithm is used to assemble the individual classified character subimages into optimal interpretations of an input string, taking into account both the quality of the overall segmentation and the degree to which each character subimage of the segmentation matches a character model. The system uses two different statistical language models, one based on a phrase dictionary and the other based on a simple word grammar. Hypotheses from recognition based on each language model are integrated using a decision tree classifier. Results from the application of the system to the recognition of handwritten responses on US census forms are reported.

Proceedings ArticleDOI
A.J. Elms1
09 Oct 1994
TL;DR: A novel method is described whereby hidden Markov models are used to recognise a horizontal "profile" of a word generated from an analysis of each vertical line of pixels in its image.
Abstract: The recognition of a printed word is traditionally limited by its segmentation into individual characters due to the noise introduced by facsimile transmission, photocopying, handling or ageing. In this paper, a character recogniser is described which does not require a prior segmentation of the word, and is thus suited to the recognition of noisy text images. A novel method is described whereby hidden Markov models are used to recognise a horizontal "profile" of a word generated from an analysis of each vertical line of pixels in its image.

Proceedings ArticleDOI
21 Jun 1994
TL;DR: A system for recognizing off-line, cursive, English text, guided in part by global characteristics (style) of the handwriting, and a new method for segmenting words into letters, based on minimizing a cost function are presented.
Abstract: We present a system for recognizing off-line, cursive, English text, guided in part by global characteristics (style) of the handwriting. We introduce a new method for segmenting words into letters, based on minimizing a cost function. Segmented letters are normalized with a novel algorithm that scales different parts of a letter separately removing much of the variation in the writing. We use a neural network for letter recognition and use the output of the network as posterior probabilities of letters in the word recognition process. We found that using a hidden Markov Model for word recognition is less successful than assuming an independent process for our small set of test words. In our experiments with several hundred words, written by 7 writers, 96% of the test words were correctly segmented, 52% were correctly recognized, and 70% were in the top three choices. >

Proceedings ArticleDOI
27 Jun 1994
TL;DR: A neural network based off-line recognition system for handwritten Chinese characters is presented, which is adaptable to accommodate newly encountered writing styles of a category and demonstrates an integration of neural computation and structural representation of Chinese characters.
Abstract: In this paper, a neural network based off-line recognition system for handwritten Chinese characters is presented. Seventeen character categories are handled with a recognition rate of 52.44%. Tolerance to shift, slight rotation and slight scaling of the input characters is achieved by the system. The approach demonstrates an integration of neural computation and structural representation of Chinese characters. Neural network is employed for its tolerance to inexactness and noise contamination of input patterns, while structural representation is adopted for its relevance to the construction of Chinese characters. Being a neural network based system, it is adaptable to accommodate newly encountered writing styles of a category. Moreover, an addition of new character categories do not require the removal of the established knowledge in the current system. >

Book ChapterDOI
01 Jan 1994
TL;DR: A representation technique for handwritten words is described which allows to construct word prototypes and apply them for recognition of bad quality words.
Abstract: This paper consists of three related parts. First, results of experiments on human perception which demonstrate potentials of text recognition are presented. Second, a representation technique for handwritten words is described which allows to construct word prototypes and apply them for recognition of bad quality words. Third, integration of information from multiple sources in word recognition and sentence recognition tasks is discussed. Results of experiments with cursive words and cheque amounts are also presented.

Proceedings ArticleDOI
25 Feb 1994
TL;DR: A novel approach that performs OCR without the segmentation step is developed and inexact matching and probabilistic evaluation used in the technique allow us to identify the correct word, by detecting a partial set of characters.
Abstract: Segmentation is a key step in current OCR systems. It has been estimated that half the errors in character recognition are due to segmentation. We have developed a novel approach that performs OCR without the segmentation step. The approach starts by extracting significant geometric features from the input document image of the page. Each feature then `votes' for the character that could have generated that feature. Thus, even if some of the features are occluded or lost due to degradation, the remaining features can successfully identify the character. In extreme case, the degradation may be severe enough to prevent recognition of some of the characters in a word. In such cases, we use a lexicon-based word recognition technique to resolve ambiguity. Inexact matching and probabilistic evaluation used in the technique allow us to identify the correct word, by detecting a partial set of characters. This paper first presents an overview of our segmentation-free OCR system and then focuses on the word-recognition technique. Preliminary experimental results show that this is a very promising approach.© (1994) COPYRIGHT SPIE--The International Society for Optical Engineering. Downloading of the abstract is permitted for personal use only.

Proceedings ArticleDOI
M. Gilloux1
09 Oct 1994
TL;DR: The main contribution of the method is a better estimation of the recognition score, i.e. the probability of generating the word with the HMM.
Abstract: This paper describes a method for improving handwritten word recognition by implicitly recognizing the writing style. Writing style is taken here to cover several types of distinctions between word shapes: cursive script vs. handprinted words, run-on vs. discrete words, differences in skew angle values, stability of lower and upper extensions of letters, presence or absence of loops in naturally looped letters. This method is applied in the general framework of hidden Markov models (HMM). The proposed method consists in the use of a set of models rather than of a unique model for each word. Each model in the set is associated to a certain class of writers, As a consequence, writing style is automatically and implicitly detected during recognition. But the main contribution of the method is a better estimation of the recognition score, i.e. the probability of generating the word with the HMM.

Proceedings ArticleDOI
21 Jun 1994
TL;DR: This paper presents a system for large vocabulary recognition of on-line handwritten cursive words that employs a TDNN-style network architecture which has been successfully used in the speech recognition domain.
Abstract: This paper presents a system for large vocabulary recognition of on-line handwritten cursive words. The system first uses a filtering module, based on simple letter features, to quickly reduce a large reference dictionary to a smaller number of candidates; the reduced lexicon along with the original input is subsequently fed to a recognition module. In order to exploit the sequential nature of the temporal data, we employ a TDNN-style network architecture which has been successfully used in the speech recognition domain. Explicit segmentation of the input words into characters is avoided by using a sliding window concept where the input word representation (a set of frames) is presented to the neural network-based recognizer sequentially. The outputs of the recognition module are collected and converted into a string of characters that can be matched with the candidate words. A description of the complete system and its components is given. >

Journal ArticleDOI
TL;DR: A new representation technique for handwritten words or other types of line images based on mapping of line segments approximating the word into a special feature space called holograph is described.

Book ChapterDOI
01 Jan 1994
TL;DR: Investigation of the second and third steps are reported, especially the use of multiple neural-nets to go from nonsegmented sequences of features to letters, indicating that the procedure is robust and promising.
Abstract: The task of recognizing handwritten English language text in off-line manner is described in terms of four interacting steps: from pixels to features, from features to letters, from letters to words and acceptance of words in terms of context We report on investigations of the second and third steps, especially the use of multiple neural-nets to go from nonsegmented sequences of features to letters Indications are that our procedure is robust and promising

Proceedings ArticleDOI
26 Jun 1994
TL;DR: A survey of applications of fuzzy set theory to handwriting recognition to describe the fuzzy k-nearest neighbor algorithm for handwritten word recognition, the fuzzy logic for handwritten street number location, and the fuzzy integral for handwritten character recognition.
Abstract: A survey of applications of fuzzy set theory to handwriting recognition is given. Three specific applications are described: 1) the fuzzy k-nearest neighbor algorithm for handwritten word recognition; 2) the fuzzy logic for handwritten street number location; and 3) the fuzzy integral for handwritten character recognition. New research utilizing fuzzy integrals in dynamic programming based handwritten word recognition is also described. >

Proceedings ArticleDOI
27 Jun 1994
TL;DR: The proposed learning is based on the dynamics of a human's hand mechanism and has been realized on a neural network for handwritten character recognition, and the recognition rates exceed those by a conventional statistical method.
Abstract: This paper proposes a new efficient learning of a neural network for handwritten character recognition. Like human learning, the proposed learning acquires excellent recognition ability for unknown character patterns only from a small number of typical character patterns. The proposed learning is based on the dynamics of a human's hand mechanism and has been realized on a neural network. The recognition rates exceed those by a conventional statistical method. >

Proceedings Article
01 Jan 1994

Proceedings ArticleDOI
03 Aug 1994
TL;DR: It is the hypothesis of the approach that handwritten character recognition is a pattern recognition problem and that there exists a set of unique features in the data which can be used for classification.
Abstract: In this research we address the problem of recognition of isolated handwritten characters. Handwritten character recognition has been a topic of research for a long period of time. It has been argued that this problem is difficult to model using classical modeling techniques, and that neural networks offer a new perspective to approaching this problem. This paper outlines the experimental evidence we have compiled while investigating possible approaches to handwritten character recognition. It is the hypothesis of our approach that handwritten character recognition is a pattern recognition problem and that there exists a set of unique features in the data which can be used for classification.


Proceedings ArticleDOI
09 Oct 1994
TL;DR: A new method on line level separation of Thai words is proposed by using topological properties of strokes and other Thai character's structural features decision trees are constructed to classify 4 groups of characters.
Abstract: This paper proposes a recognition method of off-line handwritten Thai characters from word scripts. Firstly a new method on line level separation of Thai words are proposed. Loop structure is used to classify 4 groups of characters. Secondly by using topological properties of strokes and other Thai character's structural features decision trees are constructed. Finally a recognition experiment is presented in which 100 copies of handwritten Thai words written by 10 persons are tested. Recognition rate is 99.0% and recognition time is 0.5 second per character.

Proceedings ArticleDOI
13 Nov 1994
TL;DR: This work presents a new off-line word recognition system that is able to recognise unconstrained handwritten words from their grey-scale images, and is based on structural and relational information in the handwritten word.
Abstract: We present a new off-line word recognition system that is able to recognise unconstrained handwritten words from their grey-scale images, and is based on structural and relational information in the handwritten word. We use Gabor filters to extract features from the words, and then use an evidence-based approach for word classification. A solution to the Gabor filter parameter estimation problem is given, enabling the Gabor filter to be automatically tuned to the word image properties. Our experiments show that the proposed method achieves reasonably high recognition rates compared to standard classification methods. >