Recognition of the Hungarian Fingerspelling Alphabet using Convolutional Neural Network based on Depth Data
Bence Danko, Gábor Kertész · 2018
The aim of this paper is to introduce that Convolutional Neural Networks based on depth data can be applied to recognize the signs of the Hungarian fingerspelling alphabet. After a comparison with end-to-end style RGB images, depth data is chosen as input. With the method recommended in this paper, a 94% classification accuracy was measured for the test subjects.