Image retrieval from video streams databases using similarity of clustering histogram

Abstract EN

The recent system for image retrieval based on histogram of clustering idea which considers the likeness among database of images is suggested.

Firstly, the space of image's feature is compressed using Haar transform.

Secondly points of interest were detected from wavelet image, and then those points of interest was descriptor using SURF descriptor Thirdly, the clustering algorithm of moving k-means is employed for features cluster that resulted from SURF descriptor and then a histogram was built from the cluster's values.

The suggested procedure is experimented on different database.

The outcome of experimental shows that suggested procedure is reliable, fast and active for retrieving of an image from database based on histogram than FAST detection of corner that depend on image features