Affine non-local means image denoising

This work presents an extension of the Non-Local Means denoising method, that effectively exploits the affine invariant self-similarities present in images of real scenes. Our method provides a better image denoising result by grounding on the fact that in many occasions similar patches exist in the...

ver descrição completa

Detalhes bibliográficos
Autores: Fedorov, Vadim, Ballester, Coloma
Formato: artículo
Estado:Versión aceptada para publicación
Fecha de publicación:2017
País:España
Recursos:Universitat Pompeu Fabra
Repositorio:Repositorio Digital de la UPF
OAI Identifier:oai:repositori.upf.edu:10230/37095
Acesso em linha:http://hdl.handle.net/10230/37095
http://dx.doi.org/10.1109/TIP.2017.2681421
Access Level:acceso abierto
Palavra-chave:Image denoising
Patch-based method
Patch similarity
Affine invariance
Descrição
Resumo:This work presents an extension of the Non-Local Means denoising method, that effectively exploits the affine invariant self-similarities present in images of real scenes. Our method provides a better image denoising result by grounding on the fact that in many occasions similar patches exist in the image but have undergone a transformation. The proposal uses an affine invariant patch similarity measure that performs an appropriate patch comparison by automatically and intrinsically adapting the size and shape of the patches. As a result, more similar patches are found and appropriately used. We show that this image denoising method achieves top-tier performance in terms of PSNR, outperforming consistently the results of the regular Non-Local Means, and that it provides state-of-the-art qualitative results.