Modified encoding of the significance maps in image compression with wavelets

In image compression based on transforms, the current approach to encode the position of significant coefficients consists in creating significance maps. In this paper, we propose an improvement of this approach for the case of image multiresolution analysis with wavelets. Our solution makes use of the correlation existing between the different types of details on the same level of decomposition. With this solution, the budget of bits needed to encode the significance maps is reduced by 33%, without a loss in compression quality.