Original title: Motion Analysis of Czech Sign Language Alphabet CNN based Classifier via Optical Flow
Authors: Krejsa, Jiří ; Věchet, Stanislav ; Šnajder, J.
Document type: Papers
Conference/Event: Engineering mechanics 2024 /30./, Milovy (CZ), 20240514
Year: 2024
Language: eng
Abstract: Single hand sign language alphabet letters can be successfully detected and classified from a single still image using convolution neural networks (CNN). Czech variant of the alphabet uses specific\nmotions to add diacritics to the letters. The paper deals with determination of such a motion using optical flow analysis of image sequences.
Keywords: diacritics detection; gesture recognition; sign language
Host item entry: Engineering Mechanics 2024, ISBN 978-80-214-6235-9, ISSN 1805-8248

Institution: Institute of Thermomechanics AS ČR (web)
Document availability information: Fulltext is available at the institute of the Academy of Sciences.
Original record: https://hdl.handle.net/11104/0363165

Permalink: http://www.nusl.cz/ntk/nusl-669447


The record appears in these collections:
Research > Institutes ASCR > Institute of Thermomechanics
Conference materials > Papers