Hierarchical image representation for compression, filtering and normalization

Applications of pyramid data structure to image compression and image filtering are described. The compression scheme fits visual perception models, and filtering is computationally faster than filtering in the frequency domain. Image normalization is also possible.