Dimension Reduction In Image Databases Using The Logical Combinatorial Approach

The development of techniques that allow the representation space reduction is a topic which has become of special interest, since these allow us to work with less attributes and to obtain the same results. These techniques are used mainly for the treatment of big databases, in which an immense quantity of numeric data is managed, where the final objective is the data classification. This paper presents the results of the "typical testors" technique, for reducing dimensions, and the "k-means" technique, for carrying out the object classification. The main objective of making a reduction of the representation space with typical testors is to facilitate the classification, that is, the classification is a way to check the efficiency of typical testors.