CSDL Home IEEE Transactions on Pattern Analysis & Machine Intelligence 1980 vol.2 Issue No.01 - January
Issue No.01 - January (1980 vol.2)
H. C. Du , Institute of Applied Mathematics, National Tsing Hua University, Hsinchu, Taiwan; Department of Computer Sciences, University of Washington, Seattle, WA 98105.
R. C. T. Lee , Institute of Computer and Decision Sciences, National Tsing Hua University, Hsinchu, Taiwan.
In this paper, we extend the binary Gray code to symbolic Gray code. We then show that this symbolic Gray code can be used as a multikey hashing function for storing symbolic records. The record stored at location k and the record stored at location k + 1 will be nearest neighbors if this hashing function is used. Thus, this symbolic Gray code hashing function exhibits some kind of clustering property which will group similar records together. This property is a desirable property for designing nearest neighbor searching (also called best match searching) systems. There are many other interesting properties of this hashing function. For instance, there exists an address-to-key transformation which can be used to determine the record stored at certain location k if this hashing function is used. Besides, if there are totally M records, we only have to reserve exactly M locations; there are no collisions and wasting of memory storage. At the end of this paper, it is shown that the resulting file exhibits the multiple-attribute tree structure.
H. C. Du, R. C. T. Lee, "Symbolic Gray Code as a Multikey Hashing Function", IEEE Transactions on Pattern Analysis & Machine Intelligence, vol.2, no. 1, pp. 83-90, January 1980, doi:10.1109/TPAMI.1980.4766976