Sign language, for deaf-impaired people, plays an important role in communication. In this paper, we devise a Taiwan Sign Language recognition system. We use the Kinect2 sensor to get data from 94 sign morphemes shown once by 4 people, and extract hand shape features and trajectory features from depth images and joints of the body skeleton. Finally, we have each sign morpheme dictionary trained by label consistent K-SVD (LC-KSVD) sparse coding algorithm for recognition. Experiments show our system performs well and the accuracy achieves 99.47% in close test.