Compressing Images Using Contextual Interpolator Parameterization

The paper proposes a parameterized contextual image interpolator. The parameters of the interpolator are estimated at each point of the image using the resampled images. The estimated parameters are used to interpolate the image pixels of the required resolution. The proposed parameterization reduces the number of adjustable coefficients of the context interpolator, reduces the complexity, and increases the quality of the estimation of these coefficients. The proposed interpolator is built into the compression method based on the representation of the image in the form of a quadtree. The specified embedding allows to significantly increase the compression ratio of the compression method. Computational experiments in natural digital images confirm the gain of up to nine percent in compression ratio.