Mexican sign language recognition using movement sensor

Sign language is the basic communication method between deaf people. A translator is usually needed when an ordinary person wants to communicate with a deaf one. This paper presents a system for sign language recognition of Mexican sign language to Spanish oral language. In sign language recognition, there are important challenges as to develop methods that solve signer-independent continuous sign problems. This system uses one movement sensor of image for capturing depth images and obtaining information from the skeleton of the human hand than tracking movements in the SRL. Data of this information obtained by the movement sensor. Classification and pattern recognition algorithms as random forests, decision trees and artificial neural networks process them. No gloves, special clothing or accessories are required to work it. Experimental results demonstrate than the proposed system achieves recognition Mexican signs language to letters of Spanish alphabet with an accuracy of 76.19% for sign language recognition. This system is open for future extension.