Word Database Research Articles

Creation and Analysis of (Agriculturally) Speech Database for Uttarakhand

This study is aimed to develop speech recognition database and investigate acoustic characteristics of Hindi/Garhwali vernacular used in the Garhwal region of Uttarakhand, India, ultimately to develop a spoken language interface to reap the benefits of Information and Communication Technology by farmers. A speech database of words and sentences has been created with emphasis on agricultural commodities and acoustic properties were studied. The outcome of this study will be useful for implementing speech interfaces that will be very helpful to farmers as well as people of Uttarakhand.

HamNoSys to SiGML Conversion System for Sign Language Automation

Sign language is a visual gesture language. Deaf people use sign language as their main source for communication. They use their hands, face and arms for conveying thoughts and meanings. Sign language is not having any written form and generating signs using video approach is very expensive, so, sign animation is the best solution. Research on Indian Sign Language (ISL) Automation is limited because of lack of Indian Sign Language knowledge and the unavailability of such tools which provide any education for Indian Sign Language. An automation system is designed in which HamNoSys is generated corresponding to the ISL words. HamNoSys is further converted into XML form known as SiGML. A virtual human in tool JA SiGML URL APP takes input as SiGML and generates signs corresponding to the given words. The existing tool available for generation of SiGML is eSIGNEditor which contains a database of words for American and British Sign Language while sign language in India is different from that in America and Britain. So a system is designed to convert the HamNoSys to SiGML, so as to process it further to the SiGML player to animate the signs for ISL words.

Construction of Semantic Map of Homophones for Achieving Desired Performance for Searching Homophones for a Given English Word from a Large Database of English Words

map is a complex network of words or phrases which are related in some way. To search related words or phrases for a given English word from a large vocabulary of English language is a time consuming process for a computer. Each time to search related set of words or phrases requires massive processing of comparison and if the entire database is large enough it is required to implement some mechanism that makes searching efficient and fully utilizing the computing power. The words and phrases which are related may be synonyms, antonyms or homophones meaning having similar pronunciation with different spellings and different meanings. The researcher has made an effort to prepare such a network of homophones so that when a search for homophones is required for a given English word, the fast retrieval of result in form of set of homophones can be possible. To determine about the words whether they are homophones or not we require phonetic algorithms for phonetic similarity between words. Similar to indexing a mechanism is derived using an algorithm that is one time processing to prepare a semantic map and then to retrieve set of homophones from this semantic map of homophones. This one time processing for constructing a semantic map is also somehow time consuming processing but once it is constructed searching becomes more efficient.

Improving Handwritten Arabic Character Recognition by Modeling Human Handwriting Distortions

Handwritten Arabic character recognition systems face several challenges, including the unlimited variation in human handwriting and the unavailability of large public databases of handwritten characters and words. The use of synthetic data for training and testing handwritten character recognition systems is one of the possible solutions to provide several variations for these characters and to overcome the lack of large databases. While this can be using arbitrary distortions, such as image noise and randomized affine transformations, such distortions are not realistic. In this work, we model real distortions in handwriting using real handwritten Arabic character examples and then use these distortion models to synthesize handwritten examples that are more realistic. We show that the use of our proposed approach leads to significant improvements across different machine-learning classification algorithms.

A Novel Approach to Text Dependent Writer Identification of Kannada Handwriting

In this paper a text dependent writer identification method based on Kannada handwriting is proposed. The database of Kannada handwritten words collected from 25 writers is intended to provide training and testing sets for Kannada writer identification research, because there is no well-known database containing Kannada handwritten words available for writer identification problem. A feature vector consisting of directional multi-resolution spatial features based on Radon Transform and Discrete Cosine Transform and structural features such as aspect ratio and on-pixel ratio are extracted from word images. The novel approach of combining the features of two or more words is proposed and performance analysis is done through nearest neighbor classifier with modified 5-fold cross-validation. An average identification accuracy of 93.2582% is achieved by using only the single words, and accuracy of 100% is achieved using the combination of features of three or more words. The writer identification results show that feature vectors extracted from longer words, words having more structural variation and combination of features of two or more words have higher impact on writer identification.

SIFT descriptors for Arabic handwriting recognition

In this paper, a new recognition system for off-line Arabic handwritten words is presented. The proposed system uses for the first time scale invariant feature transform SIFT descriptors to extract features from Arabic handwritten words. It is also based on keypoints matching for achieving classification. In our method, word's images are divided vertically into five frames; from each frame a set of keypoints using SIFT descriptors is extracted and compared to classes' models which includes the most discriminating keypoints extracted from words samples of the same class. Finally, the recognition process was achieved through a keypoints matching procedure, using the Euclidean distance. The paper presents also a new large Arabic handwritten word database. This database provides a new framework for benchmarking and gives a new freely available handwritten word dataset. Several tests have been performed using our new database and the well known IFN/ENIT database for comparison purposes. The reported results show the robustness and efficiency of the proposed approach.

A database of 629 English compound words: ratings of familiarity, lexeme meaning dominance, semantic transparency, age of acquisition, imageability, and sensory experience.

Since the work of Taft and Forster (1976), a growing literature has examined how English compound words are recognized and organized in the mental lexicon. Much of this research has focused on whether compound words are decomposed during recognition by manipulating the word frequencies of their lexemes. However, many variables may impact morphological processing, including relational semantic variables such as semantic transparency, as well as additional form-related and semantic variables. In the present study, ratings were collected on 629 English compound words for six variables [familiarity, age of acquisition (AoA), semantic transparency, lexeme meaning dominance (LMD), imageability, and sensory experience ratings (SER)]. All of the compound words selected for this study are contained within the English Lexicon Project (Balota et al., 2007), which made it possible to use a regression approach to examine the predictive power of these variables for lexical decision and word naming performance. Analyses indicated that familiarity, AoA, imageability, and SER were all significant predictors of both lexical decision and word naming performance when they were added separately to a model containing the length and frequency of the compounds, as well as the lexeme frequencies. In addition, rated semantic transparency also predicted lexical decision performance. The database of English compound words should be beneficial to word recognition researchers who are interested in selecting items for experiments on compound words, and it will also allow researchers to conduct further analyses using the available data combined with word recognition times included in the English Lexicon Project.

A Dempster–Shafer Theory based combination of handwriting recognition systems with multiple rejection strategies

Dempster–Shafer Theory (DST) is particularly efficient in combining multiple information sources providing incomplete, imprecise, biased, and conflictive knowledge. In this work, we focused on the improvement of the accuracy rate and the reliability of a HMM based handwriting recognition system, by the use of Dempster–Shafer Theory (DST). The system proceeds in two steps: First, an evidential combination method is proposed to finely combine the probabilistic outputs of the HMM classifiers. Second, a global post-processing module is proposed to improve the reliability of the system thanks to a set of acceptance/rejection decision strategies. In the end, an alternative treatment of the rejected samples is proposed using multi-stream HMM to improve the word recognition rate as well as the reliability of the recognition system, while not causing significant delays in the recognition process. Experiments carried out on two publically available word databases (RIMES for Latin script and IFN/ENIT for Arabic script) show the benefit of the proposed strategies.

Arabic word descriptor for handwritten word indexing and lexicon reduction

Word recognition systems use a lexicon to guide the recognition process in order to improve the recognition rate. However, as the lexicon grows, the computation time increases. In this paper, we present the Arabic word descriptor (AWD) for Arabic word shape indexing and lexicon reduction in handwritten documents. It is formed in two stages. First, the structural descriptor (SD) is computed for each connected component (CC) of the word image. It describes the CC shape using the bag-of-words model, where each visual word represents a different local shape structure, extracted from the image with filters of different patterns and scales. Then, the AWD is formed by sorting and normalizing the SDs. This emphasizes the symbolic features of Arabic words, such as subwords and diacritics, without performing layout segmentation. In the context of lexicon reduction, the AWD is used to index a reference database. Given a query image, the reduced lexicon is obtained from the labels of the first entries in the indexed database. This framework has been tested on Arabic word databases. It has a low computational overhead, while providing a compact descriptor, with state-of-the-art results for lexicon reduction on the Ibn Sina and IFN/ENIT databases.

Are there signs of change in gendered language use in children’s early reading vocabulary?

In this study we investigate the relative frequencies of female and male terms in early reading material for children using the Children’s Printed Word Database as a resource. As roles of females and males have changed over time it is of interest to see if there has been a corresponding change in representations of females and males in children’s books. We carried out analyses regarding different words related to gender. Except for nouns referring to relatives, we found in all word groups a preponderance of male terms. The imbalance of male and female pronouns is equivalent to that reported by Carroll, Davies, and Richman (1971) in a frequency count of printed words in children’s book in the USA conducted some 40 years ago. The results are discussed in terms of gender inequality in reading materials and the development of social mores and stereotypical ideas.

Individual versus Small Group Instruction of Computer Applications: A Quantitative and Qualitative Comparison

AbstractThis study seeks to answer two questions. First, how does learning of computer applications in a group setting affect individual achievement and attitude? Second, what distinctive group processes are demonstrated by groups learning computer applications? University students were randomly assigned to work individually or as members of dyads to learn word processing, spreadsheets, and databases. Quantitative data revealed that individuals and dyads had essentially identical scores on the assignments and the performance test. There were no between group differences on the attitude measures. Qualitative data indicate that groups appeared to engage in more planning activities and to have a better conceptual understanding of the applications. Dyadic learning of computer applications is discussed as a viable method of classroom instruction.Nearly two-thirds of the educators responding to a 1985 national survey attributed changes in their teaching practices to computers (Becker, 1986). Most common was inc...

Recognizing handwritten Arabic words using grapheme segmentation and recurrent neural networks

The Arabic alphabet is used in around 27 languages, including Arabic, Persian, Kurdish, Urdu, and Jawi. Many researchers have developed systems for recognizing cursive handwritten Arabic words, using both holistic and segmentation-based approaches. This paper introduces a system that achieves high accuracy using efficient segmentation, feature extraction, and recurrent neural network (RNN). We describe a robust rule-based segmentation algorithm that uses special feature points identified in the word skeleton to segment the cursive words into graphemes. We show that careful selection from a wide range of features extracted during and after the segmentation stage produces a feature set that significantly reduces the label error. We demonstrate that using same RNN recognition engine, the segmentation approach with efficient feature extraction gives better results than a holistic approach that extracts features from raw pixels. We evaluated this segmentation approach against an improved version of the holistic system MDLSTM that won the ICDAR 2009 Arabic handwritten word recognition competition. On the IfN/ENIT database of handwritten Arabic words, the segmentation approach reduces the average label error by 18.5 %, the sequence error by 22.3 %, and the execution time by 31 %, relative to MDLSTM. This approach also has the best published accuracies on two IfN/ENIT test sets.

EHME: a new word database for research in Basque language.

This article presents EHME, the frequency dictionary of Basque structure, an online program that enables researchers in psycholinguistics to extract word and nonword stimuli, based on a broad range of statistics concerning the properties of Basque words. The database consists of 22.7 million tokens, and properties available include morphological structure frequency and word-similarity measures, apart from classical indexes: word frequency, orthographic structure, orthographic similarity, bigram and biphone frequency, and syllable-based measures. Measures are indexed at the lemma, morpheme and word level. We include reliability and validation analysis. The application is freely available, and enables the user to extract words based on concrete statistical criteria 1 , as well as to obtain statistical characteristics from a list of words

Performance enhancement of online handwritten Tamil symbol recognition with reevaluation techniques

In this article, we aim at reducing the error rate of the online Tamil symbol recognition system by employing multiple experts to reevaluate certain decisions of the primary support vector machine classifier. Motivated by the relatively high percentage of occurrence of base consonants in the script, a reevaluation technique has been proposed to correct any ambiguities arising in the base consonants. Secondly, a dynamic time-warping method is proposed to automatically extract the discriminative regions for each set of confused characters. Class-specific features derived from these regions aid in reducing the degree of confusion. Thirdly, statistics of specific features are proposed for resolving any confusions in vowel modifiers. The reevaluation approaches are tested on two databases (a) the isolated Tamil symbols in the IWFHR test set, and (b) the symbols segmented from a set of 10,000 Tamil words. The recognition rate of the isolated test symbols of the IWFHR database improves by 1.9 %. For the word database, the incorporation of the reevaluation step improves the symbol recognition rate by 3.5 % (from 88.4 to 91.9 %). This, in turn, boosts the word recognition rate by 11.9 % (from 65.0 to 76.9 %). The reduction in the word error rate has been achieved using a generic approach, without the incorporation of language models.

Name that Tune! An Introduction to Musical Information Retrieval

In the beginning, computers, as their name implies, had numerical processing capability. The development of word processing and database systems gave them verbal processing capability and of graphics cards and software spatial processing capability. Now, with the widespread availability of sound cards and MIDI files, we may claim, as well, that computers have musical processing capability.

FPGA Parallel Recognition Engine for Handwritten Arabic Words

Arabic optical character recognition (OCR) has been an active field of research for decades. Yet, it has limited application domains constrained to printed text. This is partly due to high time complexity of Arabic OCR algorithms. Accelerating these algorithms allows for applications in real-time assistive technologies, office automation, and portable devices. Modern programmable hardware devices like FPGAs have numerous logic resources that allow parallel implementations of many algorithms. In this paper, we investigate implementing the feature extraction and classification stages of handwritten Arabic words on FPGAs. We study the performance and cost of four commonly-used feature extraction techniques and neural network classifiers on images from the IFN/ENIT database of handwritten Arabic words. The most efficient feature extraction technique and the best neural network found are implemented. Multiple FPGA implementations with varying cost and performance are evaluated. An implementation that only consumes about one quarter of the FPGA resources is 20 times faster than the software implementation and is less accurate by only 2.8 %.

Is it possible to create a very large wordnet in 100 days? An evaluation

Wordnets are large-scale lexical databases of related words and concepts, useful for language-aware software applications. They have recently been built for many languages by using various approaches. The Finnish wordnet, FinnWordNet (FiWN), was created by translating the more than 200,000 word senses in the English Princeton WordNet (PWN) 3.0 in 100 days. To ensure quality, they were translated by professional translators. The direct translation approach was based on the assumption that most synsets in PWN represent language-independent real-world concepts. Thus also the semantic relations between synsets were assumed mostly language-independent, so the structure of PWN could be reused as well. This approach allowed the creation of an extensive Finnish wordnet directly aligned with PWN and also provided us with a translation relation and thus a bilingual wordnet usable as a dictionary. In this paper, we address several concerns raised with regard to our approach, many of them for the first time. We evaluate the craftsmanship of the translators by checking the spelling and translation quality, the viability of the approach by assessing the synonym quality both on the lexeme and concept level, as well as the usefulness of the resulting lexical resource both for humans and in a language-technological task. We discovered no new problems compared with those already known in PWN. As a whole, the paper contributes to the scientific discourse on what it takes to create a very large wordnet. As a side-effect of the evaluation, we extended FiWN to contain 208,645 word senses in 120,449 synsets, effectively making version 2.0 of FiWN currently the largest wordnet in the world by these statistics.

An approach to automated construction of a general-purpose lexical ontology based on Wiktionary

An approach to the design of a system of automated construction of a general-purpose lexical ontology is proposed, and the architecture of such a system is described. Wiktionary is chosen as the online dictionary because it has a large database of words with translations into many languages. The structure of the dictionary entry is considered using the English Wiktionary as an example. This structure is used to design a database for storing the retrieved information. Ontologies are an important part of knowledge management systems. Ontologies require the development of approaches and algorithms for their construction. Lexical ontologies are constructed, and the main features of two ontology databases based on the Russian and English Wiktionaries are compared. The dynamics of numerical parameters of Wiktionaries and general-purpose lexical ontologies for 2010---2012 constructed by the authors is analyzed.

Handwriting word recognition using windowed Bernoulli HMMs

Hidden Markov Models (HMMs) are now widely used for off-line handwriting recognition in many languages. As in speech recognition, they are usually built from shared, embedded HMMs at symbol level, where state-conditional probability density functions in each HMM are modeled with Gaussian mixtures. In contrast to speech recognition, however, it is unclear which kind of features should be used and, indeed, very different features sets are in use today. Among them, we have recently proposed to directly use columns of raw, binary image pixels, which are directly fed into embedded Bernoulli (mixture) HMMs, that is, embedded HMMs in which the emission probabilities are modeled with Bernoulli mixtures. The idea is to by-pass feature extraction and to ensure that no discriminative information is filtered out during feature extraction, which in some sense is integrated into the recognition model. In this work, column bit vectors are extended by means of a sliding window of adequate width to better capture image context at each horizontal position of the word image. Using these windowed Bernoulli mixture HMMs, good results are reported on the well-known IAM and RIMES databases of Latin script, and in particular, state-of-the-art results are provided on the IfN/ENIT database of Arabic handwritten words.

Toward construction of semantic transparency database of two-kanji compound words (1)

Toward construction of semantic transparency database of two-kanji compound words (1)