Publication detail

Classification of Czech Sign Language Alphabet Diacritics via LSTM

ŠNAJDER, J. KREJSA, J.

Original Title

Classification of Czech Sign Language Alphabet Diacritics via LSTM

Type

conference paper

Language

English

Original Abstract

The paper presents the classification of image sequences of the single-handed Czech sign language alphabet diacritics. Since diacritics are expressed by the motion of the hand, the classification is performed by the Long Short-Term Memory recurrent neural network. Annotation of the dataset is done by the MediaPipe framework, and the neural network is constructed with the TensorFlow computational library. The paper describes the proposed method's flow, data acquisition, preprocessing, and training. Obtained results consist of the validation dataset's classification success rate and testing on whole signed words and sentences. The overall success rate was around 88%.

Keywords

Czech sign language; classification; MediaPipe; image sequence; Long Short-Term Memory

Authors

ŠNAJDER, J.; KREJSA, J.

Released

9. 12. 2022

Publisher

Institute of Electrical and Electronics Engineers Inc.

Location

Pilsen

ISBN

978-1-6654-1039-7

Book

2022 20th International Conference on Mechatronics - Mechatronika (ME)

Edition

1st

Pages from

178

Pages to

182

Pages count

5

URL

https://ieeexplore.ieee.org/document/9983436/

BibTex

@inproceedings{BUT182996,
  author="Jan {Šnajder} and Jiří {Krejsa}",
  title="Classification of Czech Sign Language Alphabet Diacritics via LSTM",
  booktitle="2022 20th International Conference on Mechatronics - Mechatronika (ME)",
  year="2022",
  series="1st",
  pages="178--182",
  publisher="Institute of Electrical and Electronics Engineers Inc.",
  address="Pilsen",
  doi="10.1109/ME54704.2022.9983436",
  isbn="978-1-6654-1039-7",
  url="https://ieeexplore.ieee.org/document/9983436/"
}