Missing Data Correction in Still Images Using Multi-Resolution Analysis

Blerim Mustafa, Vladimir Trajkovik, Danco Davcev


This paper proposes an improvement to the texture generation – based image in-painting algorithm, using multi-resolution analysis. Instead of optimizing the size of the neighborhood window of the synthesized pixel and the size of the test image, the approach described in this paper applies the same (minimal) window and test image size to different sub-bands of the discrete wavelet transformed (DWT) image. This way the execution complexity is kept lowest possible, while still obtaining the same visual quality as by applying the optimal parameters in the spatial domain.

Full Text:


DOI: https://doi.org/10.2498/cit.1000677

Creative Commons License
This work is licensed under a Creative Commons Attribution-NoDerivatives 4.0 International License.

Crossref Similarity Check logo

Crossref logologo_doaj

 Hrvatski arhiv weba logo