We present a number of methodological recommendations concerning the online evaluation of avatars for text-to-sign translation, focusing on the structure, format and length of the questionnaire, as well as methods for eliciting and faithfully transcribing responses.
LINK
Communication between healthcare professionals and deaf patients has been particularly challenging during the COVID-19 pandemic. We have explored the possibility to automatically translate phrases that are frequently used in the diagnosis and treatment of hospital patients, in particular phrases related to COVID-19, from Dutch or English to Dutch Sign Language (NGT). The prototype system we developed displays translations either by means of pre-recorded videos featuring a deaf human signer (for a limited number of sentences) or by means of animations featuring a computer-generated signing avatar (for a larger, though still restricted number of sentences). We evaluated the comprehensibility of the signing avatar, as compared to the human signer. We found that, while individual signs are recognized correctly when signed by the avatar almost as frequently as when signed by a human, sentence comprehension rates and clarity scores for the avatar are substantially lower than for the human signer. We identify a number of concrete limitations of the JASigning avatar engine that underlies our system. Namely, the engine currently does not offer sufficient control over mouth shapes, the relative speed and intensity of signs in a sentence (prosody), and transitions between signs. These limitations need to be overcome in future work for the engine to become usable in practice.
DOCUMENT
In this paper we describe our work in progress on the development of a set of criteria to predict text difficulty in Sign Language of the Netherlands (NGT). These texts are used in a four year bachelor program, which is being brought in line with the Common European Framework of Reference for Languages (Council of Europe, 2001). Production and interaction proficiency are assessed through the NGT Functional Assessment instrument, adapted from the Sign Language Proficiency Interview (Caccamise & Samar, 2009). With this test we were able to determine that after one year of NGT-study students produce NGT at CEFR-level A2, after two years they sign at level B1, and after four years they are proficient in NGT on CEFR-level B2. As a result of that we were able to identify NGT texts that were matched to the level of students at certain stages in their studies with a CEFR-level. These texts were then analysed for sign familiarity, morpheme-sign rate, use of space and use of non-manual signals. All of these elements appear to be relevant for the determination of a good alignment between the difficulty of NGT signed texts and the targeted CEFR level, although only the morpheme-sign rate appears to be a decisive indicator
DOCUMENT
Een schoone ende wonderlijcke prognosticate (1560) is one of many Dutch texts dealing with the trickster Till Eulenspiegel, known as "Tijl Uilenspiegel" in the Low Countries and "Owlglass" in England. The poem differs from most Eulenspiegel literature in two key respects. First, it treats the figure as a narrator rather than a character, and second, it seems designed for performance rather than simple recital. We offer here an English translation of this remarkable piece, lightly annotated throughout.
DOCUMENT
During the medieval and early modern periods, the mock sermon was one of the most widely staged festive rituals. There are records of its performance in most European countries and cultures. But despite its clear popularity in England, few of these texts are extant in English. Fewer still have been translated in their entirety. To remedy this situation, we provide here a translation of the sixteenth century Dutch poem Spotsermoen over Sint Niemand. This piece embodies most of the key features of the mock sermon; the text's rhetorical strategies, its content, and the details it reveals about its performance are all typical of the genre. Thus, the poem is offered as a specimen of the mock sermon's overall conventions and forms.
DOCUMENT
Posterpresentatie op conferentie Introductie: De afgelopen twee decennia is er geen communis opinio over de vraag of er sprake is van regionale factoren bij de constituentvolgorde in de Nederlandse Gebarentaal (NGT) (Coerts, 1994; Crasborn & De Wit, 2005; Vermeerbergen, Nijen Twilhaar & Van Herreweghe, 2013). Dit onderzoek levert een bijdrage aan de discussie over constituent-volgorde in het algemeen, doordat werd onderzocht of constituentvolgorde in NGT verschillend is voor eerste- en tweedetaalverwervers van de NGT. Methode: Er namen tien participanten (studenten NGT aan de Hogeschool Utrecht) deel: vijf dove moedertaalgebruikers en vijf horende studenten (leeftijd 19-30 jaar). Productie werd uitgelokt van drie verschillende typen enkelvoudige zinnen aan de hand van plaatjes (Volterra et al., 1984): 6 zinnen met omkeerbare constituenten, 6 zinnen met niet-omkeerbare constituenten en 6 locatieve zinnen. Alle 180 geproduceerde zinnen werden getranscribeerd en syntactisch (Subject, Object, Verbum) geanalyseerd. Resultaten: De moedertaalgebruikers gaven duidelijk de voorkeur aan de volgorde SVO of SVOV bij omkeerbare zinnen en SOV voor niet-omkeerbare zinnen. Voor de locatieve zinnen werd veel variatie gevonden, met een lichte voorkeur voor OSV. De T2-verwervers vertoonden veel variatie in de productie van omkeerbare zinnen zonder duidelijk patroon. Voor de niet-omkeerbare zinnen werd een lichte voorkeur voor SOV gevonden en wederom variatie. De loca-tieve zinnen werden voornamelijk uitgedrukt in de volgorde OSV. Conclusie: Deze studie toont aan dat de constituentvolgorde van moedertaalgebruikers en T2-verwervers verschillen. Opvallend is dat de T2-verwervers de neiging hebben de regel voor locatieve zinnen ‘Het grootste element voorop in de zin’ te overgeneraliseren naar het type niet-omkeerbare zin.
MULTIFILE
Growing research in sign language recognition, generation, and translation AI has been accompanied by calls for ethical development of such technologies. While these works are crucial to helping individual researchers do better, there is a notable lack of discussion of systemic biases or analysis of rhetoric that shape the research questions and methods in the field, especially as it remains dominated by hearing non-signing researchers. Therefore, we conduct a systematic review of 101 recent papers in sign language AI. Our analysis identifies significant biases in the current state of sign language AI research, including an overfocus on addressing perceived communication barriers, a lack of use of representative datasets, use of annotations lacking linguistic foundations, and development of methods that build on flawed models. We take the position that the field lacks meaningful input from Deaf stakeholders, and is instead driven by what decisions are the most convenient or perceived as important to hearing researchers. We end with a call to action: the field must make space for Deaf researchers to lead the conversation in sign language AI.
LINK
BACKGROUND: The design and manufacturing of effective foot orthoses is a complex multidisciplinary problem involving biomedical and gait pattern aspects, technical material and geometric design elements as well as psychological and social contexts. This complexity contributes to the current trial-and-error and experience-based orthopedic footwear practice in which a major part of the expertise is implicit. This hampers knowledge transfer, reproducibility and innovation. OBJECTIVE/METHODS: A systematic review of literature has been performed to find evidence of explicit knowledge, quantitative guidelines and design motivations of pedorthists. RESULTS: 17 studies have been included. No consensus is found on which measurable parameters ensure proper foot and ankle functioning. Parameters suggested are: neutral foot positioning and control of rearfoot motion, maximum arch, but also tibial internal/external rotation as well as a three point force system. Also studies evaluating foot orthoses centering on the diagnosis or orthosis type find no clear guidelines for treatment or for measuring the effectiveness. CONCLUSIONS: A gap in the translation from diagnosis to a specific, customized and quantified effective orthosis design is identified. Suggested solutions are both top-down, fitting of patient data in simulations, as well as bottom-up, quantifying current practices of pedorthists in order to develop new practical guidelines and evidence-based procedures.
DOCUMENT
Background: Concepts such as participation and environment may differ across cultures. Consequently, to use a measure like the Participation and Environment Measure for Children and Youth (PEM-CY) in other than the original English-speaking contexts, cultural adaptation needs to be assured. The aim of this study was to cross-culturally translate and adapt the PEM-CY into German as it is used in Germany, Austria and Switzerland. Methods: Fifteen parents of children and adolescents with disabilities from three German speaking countries participated in three rounds of think-aloud interviews. We followed the procedure of cultural equivalence guidelines including two additional steps. Data was analyzed by content analysis using semantic, idiomatic, experiential and conceptual equivalence. Results: Results show adaptations mainly focused on experiential and conceptual equivalence, with conceptual equivalence being the most challenging to reach. Examples of experiential equivalence included adapting the examples of activities in the PEM-CY to reflect those typical in German speaking countries. Conceptual equivalence mainly addressed aspects of “involvement” and “environment” of children and adolescents and was reached through adaptations such as enhanced instructions and structures, and additional definitions. Conclusions: This study presents a cross-cultural translation and adaptation process to develop a German version of the PEM-CY that is suitable for Germany, Austria and Switzerland. Using a modified cultural adaptation process, a culturally adapted version of PEM-CY (German) is now available for research, practice and further validation.
DOCUMENT
Background: Concepts such as participation and environment may differ across cultures. Consequently, to use a measure like the Participation and Environment Measure for Children and Youth (PEM-CY) in other than the original English-speaking contexts, cultural adaptation needs to be assured. The aim of this study was to cross-culturally translate and adapt the PEM-CY into German as it is used in Germany, Austria and Switzerland. Methods: Fifteen parents of children and adolescents with disabilities from three German speaking countries participated in three rounds of think-aloud interviews. We followed the procedure of cultural equivalence guidelines including two additional steps. Data was analyzed by content analysis using semantic, idiomatic, experiential and conceptual equivalence. Results: Results show adaptations mainly focused on experiential and conceptual equivalence, with conceptual equivalence being the most challenging to reach. Examples of experiential equivalence included adapting the examples of activities in the PEM-CY to reflect those typical in German speaking countries. Conceptual equivalence mainly addressed aspects of “involvement” and “environment” of children and adolescents and was reached through adaptations such as enhanced instructions and structures, and additional definitions. Conclusions: This study presents a cross-cultural translation and adaptation process to develop a German version of the PEM-CY that is suitable for Germany, Austria and Switzerland. Using a modified cultural adaptation process, a culturally adapted version of PEM-CY (German) is now available for research, practice and further validation.
DOCUMENT
