Image coding using a wavelet based Zernike moments compression technique

We present a new method of image coding using two popular imaging tools, Zernike moments and wavelets. The main idea is that we can produce appropriate image descriptors by involving an appropriate number of moments, compressed in a form suitable to represent an image with low reconstruction error for pattern recognition applications. At this point the concept of wavelet compression is involved, which has already been discussed in many technical papers. We use an existent wavelet based compression algorithm, to compress not the 2-D image, but the resultant moment based 1-D signal. So, using this formulation we can achieve a compressed representation of the image, suitable for pattern recognition purposes and image retrieval tasks. It is very important to notice here the ability of Zernike moments to provide a very high level of image reconstruction, using the inverse wavelet transform, establishing a useful method.