A knowledge-based sign synthesis architecture

This paper presents the modules that comprise a knowledge-based sign synthesis architecture for Greek sign language (GSL). Such systems combine natural language (NL) knowledge, machine translation (MT) techniques and avatar technology in order to allow for dynamic generation of sign utterances. The NL knowledge of the system consists of a sign lexicon and a set of GSL structure rules, and is exploited in the context of typical natural language processing (NLP) procedures, which involve syntactic parsing of linguistic input as well as structure and lexicon mapping according to standard MT practices. The coding on linguistic strings which are relevant to GSL provide instructions for the motion of a virtual signer that performs the corresponding signing sequences. Dynamic synthesis of GSL linguistic units is achieved by mapping written Greek structures to GSL, based on a computational grammar of GSL and a lexicon that contains lemmas coded as features of GSL phonology. This approach allows for robust conversion of written Greek to GSL, which is an essential prerequisite for access to e-content by the community of native GSL signers. The developed system is sublanguage oriented and performs satisfactorily as regards its linguistic coverage, allowing for easy extensibility to other language domains. However, its overall performance is subject to current well known MT limitations.     

Understanding gestures with systematic variations in movement dynamics

Sign language communication includes not only lexical sign gestures but also grammatical processes which represent inflections through systematic variations in sign appearance. We present a new approach to analyse these inflections by modelling the systematic variations as parallel channels of information with independent feature sets. A Bayesian network framework is used to combine the channel outputs and infer both the basic lexical meaning and inflection categories. Experiments using a simulated vocabulary of six basic signs and five different inflections (a total of 20 distinct gestures) obtained from multiple test subjects yielded 85.0% recognition accuracy. We also propose an adaptation scheme to extend a trained system to recognize gestures from a new person by using only a small set of data from the new person. This scheme yielded 88.5% recognition accuracy for the new person while the unadapted system yielded only 52.6% accuracy.     

面向大词汇量的实时连续中国手语识别系统

目前手语识别研究的难点之一在于如何实现大词汇量的连续语句识别，针对这个问题，该文提出了几个有效的方法，包括修正转移矩阵，状态结点的捆绑，快速匹配，在搜索路径中加入词跳转的估计参数等。利用上述技术，该文实现了一个基于数据手套和位置跟踪器的大词汇量的连续的中国手语实时识别系统，对中国手语辞典中收录的5100个词以及一批连续语句作实验，实验结果表明，文中所介绍的技术在提高系统识别速度和准确率方面都很有效。     

A rule-based translation from written Spanish to Spanish Sign Language glosses

One of the aims of Assistive Technologies is to help people with disabilities to communicate with others and to provide means of access to information. As an aid to Deaf people, we present in this work a production-quality rule-based machine system for translating from Spanish to Spanish Sign Language (LSE) glosses, which is a necessary precursor to building a full machine translation system that eventually produces animation output. The system implements a transfer-based architecture from the syntactic functions of dependency analyses. A sketch of LSE is also presented. Several topics regarding translation to sign languages are addressed: the lexical gap, the bootstrapping of a bilingual lexicon, the generation of word order for topic-oriented languages, and the treatment of classifier predicates and classifier names. The system has been evaluated with an open-domain testbed, reporting a 0.30 BLEU (BiLingual Evaluation Understudy) and 42% TER (Translation Error Rate). These results show consistent improvements over a statistical machine translation baseline, and some improvements over the same system preserving the word order in the source sentence. Finally, the linguistic analysis of errors has identified some differences due to a certain degree of structural variation in LSE.     

Gaussian Process Dynamical Models for hand gesture interpretation in Sign Language

Classifying human hand gestures in the context of a Sign Language has been historically dominated by Artificial Neural Networks and Hidden Markov Model with varying degrees of success. The main objective of this paper is to introduce Gaussian Process Dynamical Model as an alternative machine learning method for hand gesture interpretation in Sign Language. In support of this proposition, the paper presents the experimental results for Gaussian Process Dynamical Model against a database of 66 hand gestures from the Malaysian Sign Language. Furthermore, the Gaussian Process Dynamical Model is tested against established Hidden Markov Model for a comparative evaluation. A discussion on why Gaussian Process Dynamical Model is superior over existing methods in Sign Language interpretation task is then presented.     

Towards an American Sign Language interface

In this paper, we present two major parts of an interface for American Sign Language (ASL) to computer applications currently under work; a hand tracker and an ASL-parser. The hand tracker extracts information about handshape, position and motion from image sequences. As an aid in this process, the signer wears a pair of gloves with colour-coded markers on joints and finger tips. We also present a computational model of American Sign Language. This model is realized in an ASL-parser which consists of a DCG-grammar and a non-lexical component that records non-manual and spatial information over an ASL-discourse.     

Collecting and evaluating the CUNY ASL corpus for research on American Sign Language animation

While there is great potential for sign language animation generation software to improve the accessibility of information for deaf individuals with low written-language literacy, the understandability of current sign language animation systems is limited. Data-driven methodologies using annotated sign language corpora encoding detailed human movement have enabled some researchers to address several key linguistic challenges in ASL generation. This article motivates and describes our current research on collecting a motion-capture corpus of American Sign Language (ASL). As an evaluation of our motion-capture configuration, calibration, and recording protocol, we have conducted several rounds of evaluation studies with native ASL signers, and we have made use of our collected data to synthesize novel animations of ASL, which have also been evaluated in experimental studies with native signers.     

Modelling and recognition of the linguistic components in American Sign Language

The manual signs in sign languages are generated and interpreted using three basic building blocks: handshape, motion, and place of articulation. When combined, these three components (together with palm orientation) uniquely determine the meaning of the manual sign. This means that the use of pattern recognition techniques that only employ a subset of these components is inappropriate for interpreting the sign or to build automatic recognizers of the language. In this paper, we define an algorithm to model these three basic components form a single video sequence of two-dimensional pictures of a sign. Recognition of these three components are then combined to determine the class of the signs in the videos. Experiments are performed on a database of (isolated) American Sign Language (ASL) signs. The results demonstrate that, using semi-automatic detection, all three components can be reliably recovered from two-dimensional video sequences, allowing for an accurate representation and recognition of the signs.     

Feature extraction in Brazilian Sign Language Recognition based on phonological structure and using RGB-D sensors

In contrast to speech recognition, whose speech features have been extensively explored in the research literature, feature extraction in Sign Language Recognition (SLR) is still a very challenging problem. In this paper we present a methodology for feature extraction in Brazilian Sign Language (BSL, or LIBRAS in Portuguese) that explores the phonological structure of the language and relies on RGB-D sensor for obtaining intensity, position and depth data. From the RGB-D images we obtain seven vision-based features. Each feature is related to one, two or three structural elements in BSL. We investigate this relation between extracted features and structural elements based on shape, movement and position of the hands. Finally we employ Support Vector Machines (SVM) to classify signs based on these features and linguistic elements. The experiments show that the attributes of these elements can be successfully recognized in terms of the features obtained from the RGB-D images, with accuracy results individually above 80% on average. The proposed feature extraction methodology and the decomposition of the signs into their phonological structure is a promising method to help expert systems designed for SLR.     

Educational resources and implementation of a Greek sign language synthesis architecture

In this paper, we present how creation and dynamic synthesis of linguistic resources of Greek Sign Language (GSL) may serve to support development and provide content to an educational multitask platform for the teaching of GSL in early elementary school classes. The presented system utilizes standard virtual character (VC) animation technologies for the synthesis of sign sequences/streams, exploiting digital linguistic resources of both lexicon and grammar of GSL. Input to the system is written Greek text, which is transformed into GSL and animated on screen. To achieve this, a syntactic parser decodes the structural patterns of written Greek and matches them into equivalent patterns of GSL, which are then signed by a VC. The adopted notation system for the representation of GSL phonology incorporated in the system’s lexical knowledge database, is Hamburg Notation System (HamNoSys). For the implementation of the virtual signer tool, the definition of the VC follows the h-anim standard and is implemented in a web browser using a standard VRML plug-in.     

汉语框架语义知识库概述

介绍了汉语框架语义知识库(Chinese FrameNet,简称CFN)的语义描述体系。CFN是一个以Fillmore的框架语义学为理论基础、以伯克利FrameNet为参照、以汉语语料事实为依据的供计算机使用的汉语词汇语义知识库,它包括框架库、句子库和词汇库三部分。经过三年的努力,目前已取得了阶段性成果。CFN构建工程量大,难度很高,目前课题组只能在有限目标下开展工作,但是,已经探索了一条可行的技术路径,为实现语义Web中的语义知识共享以及智能化、个性化的Web服务提供了基础资源。     

A novel approach for dynamic hand gesture recognition using contour-based similarity images

A novel approach is proposed for the recognition of moving hand gestures based on the representation of hand motions as contour-based similarity images (CBSIs). The CBSI was constructed by calculating the similarity between hand contours in different frames. The input CBSI was then matched with CBSIs in the database to recognize the hand gesture. The proposed continuous hand gesture recognition algorithm can simultaneously divide the continuous gestures into disjointed gestures and recognize them. No restrictive assumptions were considered for the motion of the hand between the disjointed gestures. The proposed algorithm was tested using hand gestures from American Sign Language and the results showed a recognition rate of 91.3% for disjointed gestures and 90.4% for continuous gestures. The experimental results illustrate the efficiency of the algorithm for noisy videos.     

Simultaneous spotting of signs and fingerspellings based on hierarchical conditional random fields and boostmap embeddings

A sign language consists of two types of action; signs and fingerspellings. Signs are dynamic gestures discriminated by continuous hand motions and hand configurations, while fingerspellings are a combination of continuous hand configurations. Sign language spotting is the task of detection and recognition of signs and fingerspellings in a signed utterance. The internal structures of signs and fingerspellings differ significantly. Therefore, it is difficult to spot signs and fingerspellings simultaneously. In this paper, a novel method for spotting signs and fingerspellings is proposed. It can distinguish signs, fingerspellings and non-sign patterns, and is robust to the various sizes, scales and rotations of the signer's hand. This is achieved through a hierarchical framework consisting of three steps: (1) Candidate segments of signs and fingerspellings are discriminated using a two-layer conditional random field (CRF). (2) Hand shapes of segmented signs and fingerspellings are verified using BoostMap embeddings. (3) The motions of fingerspellings are verified in order to distinguish those which have similar hand shapes and different hand motions. Experiments demonstrate that the proposed method can spot signs and fingerspellings from utterance data at rates of 83% and 78%, respectively.     

A full and efficient machine tractable dictionary for natural language processing: A revised version of the CUVOALD

A lexicon is an essential part of any natural language processing system. The size, content and format of the lexicon is crucial in determining the power and sophistication of a natural language processing system. However, a lexicon which provides comprehensive, consistent and accurate lexical information and which is in a format facilitating fast retrieval is not easily available. This paper reports on a project which aims at the development of such a lexicon. The resulting lexicon is actually the modified and extended version of the machine tractable version of the Oxford Advanced Learner's Dictionary. The modification and extension concentrate mainly on the aspects of comprehensiveness, consistency, explicitness, accuracy and the dictionary format. The modified and extended version is considered a desirable source of lexical information for any natural language processing system.Si-Qing Chen received her Ph.D. in computational linguistics (1992) from Guangzhou Institute of Foreign Languages. His research interests include natural language processing, computer-assisted language teaching, and computational lexicography. She is currently furthering her study and research at the University of Illinois at Urbana-Champaign. She has published several articles on psycholinguistics, language teaching, philosophy of language, and natural language processing in, amongst other journals, Language Learning and the CALICO Journal.Luomai Xu is an associate professor in the Department of English at Guangzhou Institute of Foreign Languages. He has been working on several projects and published several articles on natural language processing.     

Linguistic properties based on American Sign Language isolated word recognition with artificial neural networks using a sensory glove and motion tracker

Sign language (SL), which is a highly visual–spatial, linguistically complete, and natural language, is the main mode of communication among deaf people. Described in this paper are two different American Sign Language (ASL) word recognition systems developed using artificial neural networks (ANN) to translate the ASL words into English. Feature vectors of signing words taken at five time instants were used in the first system, while histograms of feature vectors of signing words were used in the second system. The systems use a sensory glove, Cyberglove™, and a Flock of Birds^® 3-D motion tracker to extract the gesture features. The finger joint angle data obtained from strain gauges in the sensory glove define the hand shape, and the data from the tracker describe the trajectory of hand movement. In both systems, the data from these devices were processed by two neural networks: a velocity network and a word recognition network. The velocity network uses hand speed to determine the duration of words. Signs are defined by feature vectors such as hand shape, hand location, orientation, movement, bounding box, and distance. The second network was used as a classifier to convert ASL signs into words based on features or histograms of these features. We trained and tested our ANN models with 60 ASL words for a different number of samples. These methods were compared with each other. Our test results show that the accuracy of recognition of these two systems is 92% and 95%, respectively.     

Verb semantics for English-Chinese translation

A common practice in operational Machine Translation (MT) and Natural Language Processing (NLP) systems is to assume that a verb has a fixed number of senses and rely on a precompiled lexicon to achieve large coverage. This paper demonstrates that this assumption is too weak to cope with the similar problems of lexical divergences between languages and unexpected uses of words that give rise to cases outside of the pre-compiled lexicon coverage. We first examine the lexical divergences between English verbs and Chinese verbs. We then focus on a specific lexical selection problem—translating Englishchange-of-state verbs into Chinese verb compounds. We show that an accurate translation depends not only on information about the participants, but also on contextual information. Therefore, selectional restrictions on verb arguments lack the necessary power for accurate lexical selection. Second, we examine verb representation theories and practices in MT systems and show that under the fixed sense assumption, the existing representation schemes are not adequate for handling these lexical divergences and extending existing verb senses to unexpected usages. We then propose a method of verb representation based on conceptual lattices which allows the similarities among different verbs in different languages to be quantitatively measured. A prototype system UNICON implements this theory and performs more accurate MT lexical selection for our chosen set of verbs. An additional lexical module for UNICON is also provided that handles sense extension.     

Automatic categorization for improving Spanish into Spanish Sign Language machine translation

This paper describes a preprocessing module for improving the performance of a Spanish into Spanish Sign Language (Lengua de Signos Española: LSE) translation system when dealing with sparse training data. This preprocessing module replaces Spanish words with associated tags. The list with Spanish words (vocabulary) and associated tags used by this module is computed automatically considering those signs that show the highest probability of being the translation of every Spanish word. This automatic tag extraction has been compared to a manual strategy achieving almost the same improvement. In this analysis, several alternatives for dealing with non-relevant words have been studied. Non-relevant words are Spanish words not assigned to any sign. The preprocessing module has been incorporated into two well-known statistical translation architectures: a phrase-based system and a Statistical Finite State Transducer (SFST). This system has been developed for a specific application domain: the renewal of Identity Documents and Driver's License. In order to evaluate the system a parallel corpus made up of 4080 Spanish sentences and their LSE translation has been used. The evaluation results revealed a significant performance improvement when including this preprocessing module. In the phrase-based system, the proposed module has given rise to an increase in BLEU (Bilingual Evaluation Understudy) from 73.8% to 81.0% and an increase in the human evaluation score from 0.64 to 0.83. In the case of SFST, BLEU increased from 70.6% to 78.4% and the human evaluation score from 0.65 to 0.82.     

Error-tolerant sign retrieval using visual features and maximum a posteriori estimation

This paper proposes an efficient error-tolerant approach to retrieving sign words from a Taiwanese Sign Language (TSL) database. This database is tagged with visual gesture features and organized as a multilist code tree. These features are defined in terms of the visual characteristics of sign gestures by which they are indexed for sign retrieval and displayed using an anthropomorphic interface. The maximum a posteriori estimation is exploited to retrieve the most likely sign word given the input feature sequence. An error-tolerant mechanism based on mutual information criterion is proposed to retrieve a sign word of interest efficiently and robustly. A user-friendly anthropomorphic interface is also developed to assist learning TSL. Several experiments were performed in an educational environment to investigate the system's retrieval accuracy. Our proposed approach outperformed a dynamic programming algorithm in its task and shows tolerance to user input errors.