@inproceedings{tan:24006:sign-lang:lrec,
  author    = {Tan, Sihan and Miyazaki, Taro and Itoyama, Katsutoshi and Nakadai, Kazuhiro},
  title     = {{SEDA}: Simple and Effective Data Augmentation for Sign Language Understanding},
  pages     = {370--375},
  editor    = {Efthimiou, Eleni and Fotinea, Stavroula-Evita and Hanke, Thomas and Hochgesang, Julie A. and Mesch, Johanna and Schulder, Marc},
  booktitle = {Proceedings of the {LREC-COLING} 2024 11th Workshop on the Representation and Processing of Sign Languages: Evaluation of Sign Language Resources},
  maintitle = {2024 Joint International Conference on Computational Linguistics, Language Resources and Evaluation ({LREC-COLING} 2024)},
  publisher = {{ELRA Language Resources Association (ELRA) and the International Committee on Computational Linguistics (ICCL)}},
  address   = {Torino, Italy},
  day       = {25},
  month     = may,
  year      = {2024},
  isbn      = {978-2-493814-30-2},
  language  = {english},
  url       = {https://www.sign-lang.uni-hamburg.de/lrec/pub/24006.html},
  abstract  = {Sign language understanding (SLU) aims to convert sign language videos into glosses that transcribe sign language word-by-word by means of another written language and generate corresponding spoken sentences, including sign language recognition (SLR) and sign language translation (SLT). SLU has been a challenging undertaking since it demands the capability of fine-grained video understanding and sequence generation. In addition, the lack of supervised training data further hinders the advancement of SLU. To narrow the modality gap between vision and language and mitigate the data scarcity problem, we propose a Simple and Effective Data Augmentation (SEDA) framework for end-to-end SLU. In particular, SEDA consists of two key components: data augmentations on both sign and text sides and multi-task learning with task-specific fine-tuning. Experimental results on RWTH-PHOENIX Weather 2014T demonstrate that our proposed SEDA framework significantly and consistently outperforms the baseline model and achieves a WER of 19.91, a BLEU score of 25.19, and a ROUGE score of 51.72, delivering competitive scores in both SLR and SLT.}
}

@inproceedings{brock:18012:sign-lang:lrec,
  author    = {Brock, Heike and Rengot, Juliette and Nakadai, Kazuhiro},
  title     = {Augmenting Sparse Corpora for Enhanced Sign Language Recognition and Generation},
  pages     = {15--22},
  editor    = {Bono, Mayumi and Efthimiou, Eleni and Fotinea, Stavroula-Evita and Hanke, Thomas and Hochgesang, Julie A. and Kristoffersen, Jette and Mesch, Johanna and Osugi, Yutaka},
  booktitle = {Proceedings of the {LREC2018} 8th Workshop on the Representation and Processing of Sign Languages: Involving the Language Community},
  maintitle = {11th International Conference on Language Resources and Evaluation ({LREC} 2018)},
  publisher = {{European Language Resources Association (ELRA)}},
  address   = {Miyazaki, Japan},
  day       = {12},
  month     = may,
  year      = {2018},
  isbn      = {979-10-95546-01-6},
  language  = {english},
  url       = {https://www.sign-lang.uni-hamburg.de/lrec/pub/18012.html},
  abstract  = {The collection of signed utterances for recognition and generation of Sign Language (SL) is a costly and labor-intensive task. As a result, SL corpora are usually considerably smaller than their spoken language or image data counterparts. This is problematic, since the accuracy and applicability of a neural network depends largely on the quality and amount of its underlying training data. Common data augmentation strategies to increase the number of available training data are usually not applicable to the spatially and temporally constrained motion sequences of a SL corpus. In this paper, we therefore discuss possible data manipulation methods on the base of a collection of motion-captured SL sentence expressions. Evaluation of differently trained network architectures shows a significant reduction of overfitting by inclusion of the augmented data. Simultaneously, the accuracy of both sign recognition and generation was improved, indicating that the proposed data augmentation methods are beneficial for constrained and sparse data sets.}
}

@inproceedings{brock-nakadai-2018-deep:lrec,
  author    = {Brock, Heike and Nakadai, Kazuhiro},
  title     = {Deep {JSLC}: A Multimodal Corpus Collection for Data-driven Generation of {J}apanese {S}ign {L}anguage Expressions},
  pages     = {4247--4252},
  editor    = {Calzolari, Nicoletta and Choukri, Khalid and Cieri, Christopher and Declerck, Thierry and Goggi, Sara and Hasida, Koiti and Isahara, Hitoshi and Maegaard, Bente and Mariani, Joseph and Mazo,  H{\'e}l{\`e}ne and Moreno, Asuncion and Odijk, Jan and Piperidis, Stelios and Tokunaga, Takenobu},
  booktitle = {11th International Conference on Language Resources and Evaluation ({LREC} 2018)},
  publisher = {{European Language Resources Association (ELRA)}},
  address   = {Miyazaki, Japan},
  day       = {7--12},
  month     = may,
  year      = {2018},
  isbn      = {979-10-95546-00-9},
  language  = {english},
  url       = {https://aclanthology.org/L18-1670},
  abstract  = {The three-dimensional visualization of spoken or written information in Sign Language (SL) is considered a potential tool for better inclusion of deaf or hard of hearing individuals with low literacy skills. However, conventional technologies for such CG-supported data display are not able to depict all relevant features of a natural signing sequence such as facial expression, spatial references or inter-sign movement, leading to poor acceptance amongst speakers of sign language. The deployment of fully data-driven, deep sequence generation models that proved themselves powerful in speech and text applications might overcome this lack of naturalness. Therefore, we collected a corpus of continuous sentence utterances in Japanese Sign Language (JSL) applicable to the learning of deep neural network models. The presented corpus contains multimodal content information of high resolution motion capture data, video data and both visual and gloss-like mark up annotations obtained with the support of fluent JSL signers. Furthermore, all annotations were encoded under three different encoding schemes with respect to directions, intonation and non-manual information. Currently, the corpus is employed to learn first sequence-to-sequence networks where it shows the ability to train relevant language features.}
}