A Methodological and Structural Review of Hand Gesture Recognition Across Diverse Data Modalities

Overview

• Researchers have been developing Hand Gesture Recognition (HGR) systems to enable natural, efficient, and authentic human-computer interaction.
• Despite progress, automatically and precisely identifying hand gestures remains a significant challenge in computer vision.
• Recent studies have focused on specific modalities like RGB images, skeleton data, and spatiotemporal interest points.
• This paper provides a comprehensive review of HGR techniques and data modalities from 2014 to 2024, exploring advancements in sensor technology and computer vision.

Plain English Explanation

Hand gestures are a natural way for humans to communicate, especially for those who rely solely on hand gestures. Researchers have been working on developing Hand Gesture Recognition (HGR) systems to make it easier for people to interact with computers using their hands. However, accurately recognizing hand gestures automatically is still a significant challenge in computer vision.

Recent studies have focused on different ways of capturing and analyzing hand gestures, such as using RGB images, skeleton data, and spatiotemporal interest points. This paper reviews the progress made in HGR techniques and data modalities over the past decade, from 2014 to 2024, and explores how improvements in sensor technology and computer vision have contributed to this field.

Technical Explanation

The researchers reviewed over 200 articles on HGR from prominent databases, focusing on data collection, data settings, and gesture representation. They explored advancements in various data modalities, including RGB, Skeleton, Depth, Audio, EMG, EEG, and Multimodal approaches. The review assesses the efficacy of HGR systems through their recognition accuracy and identifies a gap in research on continuous gesture recognition, indicating the need for improved vision-based gesture systems.

The paper highlights the steady research progress in the field, including advancements in hand-crafted features and deep learning (DL) techniques. It also reports on the promising developments in HGR methods and the area of multimodal approaches, where multiple sensing modalities are combined to enhance gesture recognition.

Critical Analysis

The paper provides a comprehensive review of the HGR field, covering a wide range of data modalities and techniques. However, it also acknowledges the need for further research on continuous gesture recognition, as most existing systems focus on isolated gestures. This limitation highlights the challenge of developing robust and practical HGR systems that can handle real-world scenarios with continuous gestures.

Additionally, the review could have delved deeper into the specific challenges and limitations of the various data modalities and techniques discussed. For example, the paper could have explored the trade-offs between the accuracy, computational complexity, and user experience of different approaches.

Conclusion

This survey paper offers a valuable overview of the advancements in Hand Gesture Recognition (HGR) techniques and data modalities over the past decade. It highlights the steady progress made in the field, including improvements in hand-crafted features and deep learning methods, as well as the promising developments in multimodal approaches.

The paper's findings suggest that while significant progress has been made, there is still room for improvement, particularly in continuous gesture recognition. Addressing this gap could lead to more natural and efficient human-computer interaction, benefiting those who rely on hand gestures for communication. The review serves as a valuable resource for researchers and practitioners in the field of computer vision and human-computer interaction.