Hierarchical Representation of Plain Areas of Post-Interpolation Residuals for Image Compression
Mikhail V. Gashnikov · 2021 Ural Symposium on Biomedical Engineering, Radioelectronics and Information Technology (USBEREIT) · 2021
We propose an algorithm for encoding quantized post-interpolation residuals within the framework of hierarchical image compression. This coding algorithm is based on a hierarchical representation of the plain areas of quantized post-interpolation residuals to improve the coding efficiency of these areas. The proposed algorithm reorders the post-interpolation residuals to increase the size of the plain areas. We embed the proposed coding algorithm for post-interpolation residuals into a hierarchical image compression method. This method is based on interpolation the image scale levels using more resampled scale levels of the same image. The errors of this interpolation (post-interpolation residuals) are then quantized and encoded. We use the proposed algorithm to encode the quantized post-interpolation residuals of the hierarchical compression method. We perform computational experiments to study the effectiveness of the proposed algorithm for a set of natural images. We experimentally confirm that the use of the proposed coding algorithm for post-interpolation residuals makes it possible to increase the efficiency of the hierarchical method of image compression.