Academy of Taiwan Information Systems Research (ATISR)
Abstract:
Sign language plays an important role in communicate with changers that
hearing improved. However, the sign language in many countries and areas different
and auto recognition system became the research way in recent year. In this paper, we
devise a novel method for occlusion processing in Taiwan Sign Language recognition
system. Our method employs adxl345 and Kinect to extract the feature of signer. Then
the features are regulated by the dictionary of sparse coding. In final, the HMM model
and result signs are recognized from the features that corrected by our method. In
experimental result, we present the data that our employ. Then we describe closing
test result and future work.
Relation:
International Conference on Internet Studies (NETs 2015)
