.png){kind=icon}
Return
Nonlinear graph wavelet image compression
.png){kind=icon}
PRE
AI
DOI：10.1016/j.sigpro.2026.110838.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
A still image compression system that is based on a nonlinear graph signal transform is developed in this work. The transform nonlinearity stems from the use of median operators in a lifting structure. The transform generates a multiresolution subband structure and the subband coefficients are encoded using the SPIHT (Set Partitioning in Hierarchical Trees) algorithm. An efficient in-place algorithm is developed for implementing the transform for images, where the use of adjacency matrices is not needed. Extensive performance evaluation is presented and comparisons are made with the classical wavelet compression method. It will be shown that for some types of images, with edges as the primary feature, the nonlinear transform can give a lower bit-rate, for a similar perceptual quality as measured by the SSIM (Structural Similarity Index Measure).
Keywords:
Graph wavelets
Image compression
Median filter
Nonlinear filter banks
Journal
IF:
3.6
Papers:
10.0K
Citations:
1.7W
Organization
Cited Papers
No cited papers available