Singular Value Decomposition Image Compression System for Automatic Object Recognition

O.O. Vergara Villegas, R.P. Elías, and V.G. Cruz Sánchez (Mexico)


Singular Value Decomposition, Image Compression, Pattern recognition, Automatic Object Recognition.


In this paper we present a lossy approach to compress digital images, in which the decompressed images have the sufficient quality to be used for Computer Vision (CV) tasks as pattern recognition. To compress the images we use the Singular Value Decomposition (SVD) transform, which allow to refactoring a digital image in three matrices. The use of the resulting singular values of such refactoring allows us to represent the image with a smaller set of values and achieve the lossy image compression process. This compression method preserves useful features of the original image to perform automatic object recognition (AOR) over the decompressed image using software of National Instruments called “Vision Builder”. Additionally we show the preliminary results of the constructed system for compression/decompression and for the AOR

Important Links:

Go Back