In this paper we evaluate the performance of Conditional Random Fields (CRF) and Hidden Markov Models when recognizing motion based gestures in sign language. We implement CRF, Hidden CRF and Latent-Dynamic CRF based systems and compare these to a HMM based system when recognizing motion gestures and identifying inter gesture transitions. We implement a extension to the standard HMM model to develop a threshold HMM framework which is specifically designed to identify inter gesture transitions. We evaluate the performance of this system, and the different CRF systems, when recognizing gestures and identifying inter gesture transitions. ©2009 IEEE.