Advancements in Gesture Recognition: From Traditional Machine Learning to Deep Learning Innovations

Qingyang Wang

2024

Abstract

Gesture recognition is essential for creating more efficient human-computer interactions, transforming the way people communicate with and control technology. With improvements in computer performance and the development of image processing technology, researchers have begun to explore how to automatically extract useful information to achieve effective gesture recognition. This paper focuses on the advancements in gesture recognition, highlighting the progression from conventional machine learning to state-of-the-art deep learning approaches. Traditional machine learning is limited by its feature dependency and offers limited accuracy but has low computational complexity and strong interpretability. Convolutional Neural Network (CNN)-based methods are characterized by automatic feature extraction, high recognition accuracy, and adaptability to complex environments, but they come with high computational demands and data dependence. Transformer-based methods excel in capturing global information and have high recognition accuracy potential but are affected by extremely high computational complexity and a vast model optimization space. In summary, each of the three gesture recognition methods has its own benefits and disadvantages, and in real-world applications, the best approach should be chosen depending on specific needs and scenarios.

Paper Citation

in Harvard Style

Wang Q. (2024). Advancements in Gesture Recognition: From Traditional Machine Learning to Deep Learning Innovations. In Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM; ISBN 978-989-758-738-2, SciTePress, pages 372-376. DOI: 10.5220/0013332100004558

in Bibtex Style

@conference{mlscm24,
author={Qingyang Wang},
title={Advancements in Gesture Recognition: From Traditional Machine Learning to Deep Learning Innovations},
booktitle={Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM},
year={2024},
pages={372-376},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013332100004558},
isbn={978-989-758-738-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 1st International Conference on Modern Logistics and Supply Chain Management - Volume 1: MLSCM
TI - Advancements in Gesture Recognition: From Traditional Machine Learning to Deep Learning Innovations
SN - 978-989-758-738-2
AU - Wang Q.
PY - 2024
SP - 372
EP - 376
DO - 10.5220/0013332100004558
PB - SciTePress