Joint denoising and decompression using CNN regularization

Mario González (1) , Javier Preciozzi (1) , Pablo Musé (1) , Andrés Almansa (2)

1 IIE - Instituto de Ingeniería Eléctrica
2 MAP5 - UMR 8145 - Mathématiques Appliquées Paris 5

Wavelet compression schemes (such as JPEG2000) lead to very specific visual artifacts due to the quantization of noisy wavelet coefficients. They have highly spatialy-correlated structure that makes it difficult to be removed with standard denoising algorithms. In this work, we propose a joint denoising and decompression method that combines a data-fitting term which takes into account the quantization process and an implicit prior contained in a stateof-the-art denoising CNN.

Domaines: Traitement du signal et de l'image [eess.SP]; Traitement des images [eess.IV]; Vision par ordinateur et reconnaissance de formes [cs.CV]; Analyse numérique [math.NA]; Machine Learning [stat.ML]; Analyse numérique [cs.NA]; Traitement du signal et de l'image [eess.SP]; Optimisation et contrôle [math.OC]

Mario González, Javier Preciozzi, Pablo Musé, Andrés Almansa. Joint denoising and decompression using CNN regularization. CVPR Workshop and Challenge on Learned Image Compression (CVPR 2018), IEEE/CVF, Jun 2018, Salt Lake City, UT, United States. ⟨hal-01825573⟩