Gesture Recognition of RGB and RGB-D Static Images Using Convolutional Neural Networks

TitleGesture Recognition of RGB and RGB-D Static Images Using Convolutional Neural Networks
Publication TypeJournal Article
Year of Publication2019
AuthorsKhari, M., A. K. Garg, R. Gonzalez-Crespo, and E. Verdú
JournalInternational Journal of Interactive Multimedia and Artificial Intelligence
ISSN1989-1660
IssueRegular Issue
Volume5
Number7
Date Published12/2019
Pagination22-27
Abstract

In this era, the interaction between Human and Computers has always been a fascinating field. With the rapid development in the field of Computer Vision, gesture based recognition systems have always been an interesting and diverse topic. Though recognizing human gestures in the form of sign language is a very complex and challenging task. Recently various traditional methods were used for performing sign language recognition but achieving high accuracy is still a challenging task. This paper proposes a RGB and RGB-D static gesture recognition method by using a fine-tuned VGG19 model. The fine-tuned VGG19 model uses a feature concatenate layer of RGB and RGB-D images for increasing the accuracy of the neural network. Finally, on an American Sign Language (ASL) Recognition dataset, the authors implemented the proposed model. The authors achieved 94.8% recognition rate and compared the model with other CNN and traditional algorithms on the same dataset.

KeywordsConvolution Neural Network, Gesture Recognition, Image Processing, Sign Language
DOI10.9781/ijimai.2019.09.002
URLhttps://www.ijimai.org/journal/sites/default/files/files/2019/09/ijimai20195_7_2_pdf_18405.pdf
AttachmentSize
IJIMAI20195_7_2.pdf780.99 KB