An isolated Vietnamese Sign Language Recognition method using a fusion of Heatmap and Depth information based on Convolutional Neural Networks

In recent years, interest in sign language recognition has continuously increased. However, recognition methods for exploiting the combination of RGB and depth data are limited, especially applied to Vietnamese sign language. This paper presents an isolated Vietnamese sign language recognition method using a novel streams-enhanced 3D ConvNet. The experimental results demonstrate the superiority of the proposed method over other methods using variations from RGB, depth, and RGB-D data. The speed and accuracy of our method are better than those of previous methods.