Cyclic Linear Hidden Markov Models for Shape Classification
In classification tasks, shape descriptions, combined with matching techniques, must be robust to noise and invariant to transformations. Most of these distortions are relatively easy to handle, particularly if we represent contours by sequences. However, starting point invariance seems to be difficult to achieve. The concept of cyclic sequence, a sequence that has no initial/final point, can be of great help. We propose a new methodology to use HMMs to classify contours represented by cyclic sequences. Experimental results show that our proposal significantly outperforms other methods in the literature.
KeywordsCyclic Sequences Hidden Markov Models Shape Classification
- 5.Arica, N., Yarman-Vural, F.: A shape descriptor based on circular hidden markov model. In: International Conference on Pattern Recognition, vol. I, pp. 924–927 (2000)Google Scholar
- 9.Rabiner, L.R.: A tutorial on hidden Markov models and selected applications in speech recognition. Proc. IEEE 77(2) (1989)Google Scholar
- 10.Young, S., Odell, J., Ollason, D., Valtchev, V., Woodland, P.: The HTK Book. Cambridge University 1996 (1995)Google Scholar