Optimal trees for searching in codebook
Finding nearest neighbour of a given vector in a codebook leads to the following model of searching. In a metric space V, a vector x and a finite subset of vectors S (representing a codebook) are given. We have to find an element of S which is „nearest” to the element x. In what follows, the problem is formulated more exactly and a characterization of optimal search trees for this model of searching is given. It turns out, that balanced quasi-ternary trees are optimal search trees for the discussed problem. The result enables to speed up fording a codebook representation vector of a given acoustic vector, which is important for applications in speech recognition and synthesis.
Unable to display preview. Download preview PDF.
- 1.M. Berka, I. Kopeček. Optimality of Decision Structures of Information Systems. 16th IFIP Conference on System Modelling and Optimization, 1993, pages 653–657.Google Scholar
- 2.D.Y. Cheng, A. Gerscho. A Fast Codebook Search Algorithm for Nearest-Neighbour Pattern Matching. Proc ICASSP 86, Tokyo 1986, pages 265–268.Google Scholar
- 3.R.M. Gray. Vector Quantization. IEEE ASSP Magazine, 1, 1984, pages 4–29.Google Scholar
- 4.I. Kopeček. Speech Synthesis of Czech Language in Time Domain and Applications for Visually Impaired. Proceedings of the 2-nd SQEL Workshop, Plzeň 1997, pages 141–144.Google Scholar
- 5.D. E. Knuth. The Art of Computer Programming. Volume 3-Sorting and Searching. Addison-Wesley, 1973.Google Scholar
- 6.W. Mayeda. Graph Theory. New York, Wiley 1972.Google Scholar
- 7.J. Wiedermann. Searching. SNTL, Praha 1991 (in Czech).Google Scholar
- 8.T.P. Yunck. A Technique to Identify Nearest Neighbours. IEEE Trans. Syst. 6(76), pages 678–683.Google Scholar