Enhancing Education for Deaf People: A Systematic Review of NLP Strategies for Automatic Translation From Portuguese to Brazilian Sign Language

This research-to-practice full paper presents a comprehensive and objective account of a study that collates, analyses, and synthesizes all available evidence on a specific research question. In light of the growing demand for inclusive solutions for the deaf community, driven by the impact of the global pandemic, technologies such as Speech-to-Text (STT) and Natural Language Processing (NLP) assume a pivotal role in promoting accessibility and effective communication. A systematic literature review was conducted to analyze tools that utilize natural language processing (NLP) and speech-to-text (STT) to translate Portuguese into Brazilian Sign Language (Libras) and evaluate their effectiveness. Some tools were identified, including OpenSigns, Hand Talk, and Virtual Sign, each of which exhibited certain limitations and challenges. To evaluate the quality of the translations, metrics such as Bilingual Evaluation Understudy (BLEU) score, Word Error Rate (WER), and Character Error Rate (CER) were employed. Furthermore, the potential of the Transformer and LightConv models as promising avenues for enhancing the efficiency and accuracy of translations was under-scored. Nevertheless, considerable challenges persist, including the adaptation of cultural and grammatical nuances and the expansion of the character vocabulary to elevate translation quality.