Electromyography Based Translator of the Polish Sign Language

Noemi Kowalewska*, Przemysław Łagodziński, Marcin Grzegorzek

*Corresponding author for this work


Difficult problem is to facilitate the communication of a deaf person who communicates using sign language with a person who does not know such a way of conversation. A solution to problem may be to analyze hands movement in real time into written or spoken language. To achieve this, the experiment uses machine learning, which is used to recognize words in hand movements. It is also investigated what sensor combinations such as gyroscope, accelerometer, electromyograph give the best results. This work analyzes only a small part of all sign language words, but shows that much can be achieved by analyzing the hand movement itself. The project classified 18 words and with the proposed approach satisfactory results were obtained showing performance of 91%. It also presents problems that may arise when classifying the whole language.

Original languageEnglish
Title of host publicationITIB 2019: Information Technology in Biomedicine
EditorsEwa Pietka, Pawel Badura, Jacek Kawa, Wojciech Wieclawek
Number of pages10
PublisherSpringer, Cham
Publication date26.06.2019
ISBN (Print)978-3-030-23761-5
ISBN (Electronic)978-3-030-23762-2
Publication statusPublished - 26.06.2019
Event7th International Conference on Information Technology in Biomedicine
- Kamień Śląski, Poland
Duration: 18.06.201920.06.2019
Conference number: 229749


Dive into the research topics of 'Electromyography Based Translator of the Polish Sign Language'. Together they form a unique fingerprint.

Cite this