Author
Listed:
- Padmavathi B.
- Aarthi Elaveini M.
- Kapileswar N.
- Judy Simon
- Reshma P Vengaloor
Abstract
Gesture recognition serves as a vital interface in human-machine communication, enabling systems to interpret and respond to user commands through natural body movements, particularly hand gestures. In the development of smart environments, assistive systems, and augmented reality applications, accurate and real-time gesture interpretation is essential. However, gesture recognition faces several challenges, including variations in lighting, background complexity, hand occlusions, and the temporal dynamics of human gestures. Existing approaches primarily depends on RGB data, making them susceptible to environmental noise and fluctuations in illumination. Additionally, some existing methods are ineffective in modeling temporal dependencies, resulting in decreased recognition reliability. To address these limitations, this research proposes a novel architecture, DMT-GAFNet, designed to enhance gesture command recognition by integrating dual-modality encoding with a guided attention fusion model. The model incorporates parallel encoders for RGB and thermal streams, alongside a modality confidence estimator that dynamically weights features based on input reliability. A lightweight GRU-based temporal encoder ensures effective sequential modeling of gestures. The system was experimentally validated on a dataset combining HaGRID RGB data and Zenodo thermal data, encompassing six gesture classes and diverse visual conditions. Comparative analysis with existing deep learning models, including CNN-LSTM, MobileNetV2, ResNet18, EfficientNetB0, and VGG16, demonstrates that the proposed model outperforms these alternatives, achieving a precision of 0.9399, recall of 0.9484, F1-score of 0.9493, specificity of 0.9523, and accuracy of 97.05%. The proposed method not only achieves high classification accuracy under varying conditions but also exhibits significant potential for deployment in real-time gesture-based interaction systems.
Suggested Citation
Handle:
RePEc:gvu:iroiip:v:7:y:2025:i:2:id:785
DOI: 10.36548/jiip.2025.2.007
Download full text from publisher
Corrections
All material on this site has been provided by the respective publishers and authors. You can help correct errors and omissions. When requesting a correction, please mention this item's handle: RePEc:gvu:iroiip:v:7:y:2025:i:2:id:785. See general information about how to correct material in RePEc.
If you have authored this item and are not yet registered with RePEc, we encourage you to do it here. This allows to link your profile to this item. It also allows you to accept potential citations to this item that we are uncertain about.
We have no bibliographic references for this item. You can help adding them by using this form .
If you know of missing items citing this one, you can help us creating those links by adding the relevant references in the same way as above, for each refering item. If you are a registered author of this item, you may also want to check the "citations" tab in your RePEc Author Service profile, as there may be some citations waiting for confirmation.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: Journal Manager (email available below). General contact details of provider: https://irojournals.com/iroiip .
Please note that corrections may take a couple of weeks to filter through
the various RePEc services.