This paper presents DA3D (Data Adaptive Dual Domain Denoising), a “last step denoising” method that takes as input a noisy image and as a guide the result of any state-of-the-art denoising algorithm. The method performs frequency domain shrinkage on shape and data-adaptive patches. Unlike other dual denoising methods, DA3D doesn’t process all the image samples, which allows it to use large patches (64×64 pixels). The shape and data-adaptive patches are dynamically selected, effectively concentrating the computations on areas with more details, thus accelerating the process considerably. DA3D also reduces the staircasing artifacts sometimes present in smooth parts of the guide images. The effectiveness of DA3D is confirmed by extensive experimentation. DA3D improves the result of almost all state-of-the-art methods, and this improvement requires little additional computation time.
The preprint of the article is available here.
The supplementary material for the article is available here.
The source code for DA3D is available here (zip archive).
An online demo for DA3D is available here (user demo and password demo).