IMAGE RETRIEVAL BASED ON RADON AND WAVELET TRANSFORM
An Zhi · JOURNAL OF INFRARED AND MILLIMETER WAVES · 2008
A new kind of image retrieval algorithm based on the texture feature was presented by using the Radon and wavelet transform.Firstly,the orientation-normalized image was achieved by rotating its principal axis to horizontal direction and the Radon transform was used to deal with the orientation-normalized image.It can get the rotation-invariant Radon project of the image.Secondly,the new adoptive wavelet decomposition for the Radon project of the image was constructed according to the property of Radon transform,which had the property of shift-and scale-invariant to the project of Radon transform.Then the wavelet coefficients had the property of rotation-,translation-and scale-invariant to the project of Radon transform.Finally,the energy of every sub frequency bands was extracted to be the texture feature of image.The image retrieval experiments indicate that this method has the property of shift,scale and rotation invariance and it has a higher retrieval-rate than other methods'.