A plurality of compressed data obtained by information-compressing a plurality of data by a discrete cosine transform (DCT) are stored in a main image file. When a retrieval instruction is applied, data serving as a data retrieval key is extracted from an image database, and the extracted data is DCT-compressed by a compression processing unit. Using DC components of the resultant compressed data, reference data are generated. A search unit collates the reference data with the DC component data of the compressed data stored in the main image file and extracts data on the basis of this collation result. With this arrangement, in retrieving desired data from a plurality of data groups stored in the information-compressed state, data collation processing can be performed in the information-compressed state.

 
Web www.patentalert.com

< (none)

< Adaptive similarity searching in sequence databases

> Encoding method of members related by multiple concept or group hierarchies and identification of members in a corpus or a database that are descendants of one or more selected concepts or groups from the encoding

> (none)

~ 00001