A Vision-Based Sign Language Recognition using Statistical and Spatio-Temporal Features

DOI

10.2991/978-94-6463-196-8_21How to use a DOI?

Keywords

Sign Language Recognition (SLR); Spatio-Temporal; Analysis of variance (ANOVA)

Abstract

Those with disabilities should not be characterised primarily by their impairment in modern society; rather, it is the environment that may disable persons with disabilities. As automatic Sign Language Recognition (SLR) develops, digital technology will give more enabling settings. Many existing SLR techniques focus on the classification of static hand gestures, despite the fact that communication is a time activity, as many dynamic gestures demonstrate. As a result, temporal information obtained during the delivery of a gesture is rarely considered in SLR. The studies in this paper look at the challenge of SL gesture identification in terms of how dynamic gestures vary throughout delivery, and the goal of this research is to see how single and mixed characteristics affect a machine learning model’s classification abilities. A complex categorization task is presented with 18 frequent movements captured using a Leap Motion Controller sensor. Statistical descriptors and spatio-temporal properties are among the features derived from a 0.6 s time window. Each set’s features are compared using ANOVA F-Scores and p-values, then sorted into bins of 10 features each, up to a maximum of 250. The best statistical model chose 240 features and achieved an accuracy of 85.96%, the best spatio-temporal model chose 230 features and achieved an accuracy of 80.98%, and the best mixed-feature model chose 240 features from each set and achieved an accuracy of 86.75%. When all three sets of results are examined, the overall distribution indicates that when inputs are any number of mixed features versus any number of either of the two single sets of features, the minimum outcomes are raised.

Copyright

© 2023 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Prashant Rawat
AU  - Lalit Kane
PY  - 2023
DA  - 2023/08/10
TI  - A Vision-Based Sign Language Recognition using Statistical and Spatio-Temporal Features
BT  - Proceedings of the First International Conference on Advances in Computer Vision and Artificial Intelligence Technologies (ACVAIT 2022)
PB  - Atlantis Press
SP  - 262
EP  - 277
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6463-196-8_21
DO  - 10.2991/978-94-6463-196-8_21
ID  - Rawat2023
ER  -