Online Evaluation of Text-to-sign Translation by Deaf End Users: Some Methodological Recommendations (short paper)

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

Frozen Pretrained Transformers for Neural Sign Language Translation

One of the major challenges in sign language translation from a sign language to a spoken language is the lack of parallel corpora. Recent works have achieved promising results on the RWTH-PHOENIX-Weather 2014T dataset, which consists of over eight thousand parallel sentences between German sign language and German. However, from the perspective of neural machine translation, this is still a tiny dataset. To improve the performance of models trained on small datasets, transfer learning can be used. While this has been previously applied in sign language translation for feature extraction, to the best of our knowledge, pretrained language models have not yet been investigated. We use pretrained BERT-base and mBART-50 models to initialize our sign language video to spoken language text translation model. To mitigate overfitting, we apply the frozen pretrained transformer technique: we freeze the majority of parameters during training. Using a pretrained BERT model, we outperform a baseline trained from scratch by 1 to 2 BLEU-4. Our results show that pretrained language models can be used to improve sign language translation performance and that the self-attention patterns in BERT transfer in zero-shot to the encoder and decoder of sign language translation models.

Develop a Parallel Processing System to Recognize the Deaf and Mute Gestures and Apply it based on FPGA

In this paper, we processed an array which represents the human hand image to get the characteristics of this image. So, we used FPGA technique, and the processing operation is partitioned into three threads which is carried out in parallel. Each thread is carried out using the pipeline technique by partitioning thread into four segments. After that, we evaluated the speedup that we get in result of using the pipeline technique and the parallel threads. So, we have the possibility to design an embedded system integrated into chip (SoC), and using the mobile phones as integral devices support the software and hardware resources.

The Effect of Hearing Impairment on The Development of Communication of Poor-Hearing Children

This study aims to identifythe effect of hearing impairment on the the development of communication of poor-hearing children. Validity and reliability have been examined for the instrument and were approved to be usedin the current study.The instrument was developed by the researcher and consisted of (32) items. The Sample consisted of (50) families of the hearing impaired children. The most frequent problemswerethe child’s request for repeating the speech and were related to finding difficulties in continuing the conversation. The least common problem was that the child moved away from others. The finding showed that there were no significant statistical differences in the effect of gender of poor-hearing children on the development of communicationand in their parents’ educathional level.