Rotation Invariant Indexing For Image Using Zernike Moments and R–Tree

The Zernike moment algorithm and R-Tree algorithm are known as state of the art in the recognition of images and in the multimedia database respectively. The methods of storing the images and retrieving the similar images based on a query image automatically are the problems in the image database. This paper proposes the method to combine the Zernike moments algorithm and the R–tree algorithm in the image database. The indices of images which are retrieved from the extraction process using Zernike moments algorithm are used as the multidimensional indices to recognize the images. The multidimensional indices of Zernike moments which are stored in the R–tree are compared to the magnitudes of Zernike moments of a query image for searching the similar images. The result shows that the combination of these algorithms can be used efficiently in the image database because the recognition accuracy rate using Zernike moments algorithm is 95.20%.