| Abstract [eng] |
Ever-increasing flows of graphical information (digital images) lead to the perception that it’s necessary to search for new efficient image processing algorithms and ideas. Over the last few decades, the discrete wavelet transform, as well as wavelets themselves, has gained widespread acceptance in digital image and signal processing. Many kernels (mother wavelets) can be used for the discrete wavelet transform, for instance, Haar, Le Gall, Daubechies, etc. The discrete reversible (integer-to-integer) Le Gall wavelet transform is of particular importance. Among new solid applications of the integer Le Gall wavelet transform, there are image pattern recognition, progressive image encoding, defect localization in textural images and so on. In this work, a new original procedure for finding the discrete reversible (integer-to-integer) Le Gall wavelet transform (spectrum) of the selected image blocks is presented. The presented procedure leans upper the assumption that the discrete Le Gall wavelet spectrum of the original input image is known, and appears to be much faster than direct evaluation of Le Gall spectra for the respective image blocks. |
