Image denoising with block-matching and 3D filtering

By Kostadin Dabov, Alessandro Foi, Vladimir Katkovnik, Karen O. Egiazarian. Electronic imaging, 2006

851 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/742b3dce4a954001bb31bc7fa739c93fcbccd289

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

In Media Technology

Citations in the corpus

Learning to Restore Low-Light Images via Decomposition-and-Enhancement (2020, methodology): “Specifically, we choose one classic denoising method BM3D [11] and one recent deep learning based denoising method xDnCNN [27] to pre-/post-process the low-light images (in the test set) before/after they are processed by enhancement method LIME [20].”
Attention Guided Low-Light Image Enhancement with a Large Scale Low-Light Simulation Dataset (2019, methodology): “For Gaussian denoising, BM3D [16] and DnCNN [84] are representatives of the filter-based and deep-learning-based methods.”
Intermediate Layer Optimization for Inverse Problems using Deep Generative Models (2021, neutral): “We also compare with a standard denoising method, the BM3D algorithm (Dabov et al., 2006).”
Multiresolution Bilateral Filtering for Image Denoising (2008, neutral): “[10]–[12] develop gray-scale and color image denoising algorithms based on sparse and redundant representations over learned dictionaries , where training and denoising use the K-SVD algorithm; [13] and [14] group 2-”
Restoring an Image Taken through a Window Covered with Dirt or Rain (2013, methodology): “For the rain images, we used similar parameters, but adjusted for the fact that the images were downsampled by half: 5 × 5 for the median filter, σ s = 2 and σ r = 0.3 for the bilateral filter, and σ = 0.15 for BM3D.”
Deep Mean-Shift Priors for Image Restoration (2017, methodology): “[33] build on the ADMM algorithm and propose to replace the proximal operator of the regularizer with a denoiser such as BM3D [9] or NLM [5].”
Learning Multiscale Sparse Representations for Image and Video Restoration (2008, neutral): “Our results for denoising grayscale images outperform, for instance, works such as [10, 19, 20, 22, 35, 38].”
Generating Training Data for Denoising Real RGB Images via Camera Pipeline Simulation (2019, methodology): “The current state-of-the-art classical denoising method remains BM3D [8].”
Image Denoising by Sparse 3-D Transform-Domain Collaborative Filtering (2007, background): “It generalizes and improves our preliminary algorithm introduced in [12].”
Ask the locals: Multi-way local pooling for image recognition (2011, positive): “The usefulness of considering similar inputs for smoothing noisy data over a homogeneous sample to reduce noise without throwing out the signal has long been recognized in the image processing and denoising communities [6, 8], and has been successfully incorporated to denoising methods using sparse ”
Color Image Denoising via Sparse 3D Collaborative Filtering with Grouping Constraint in Luminance-Chrominance Space (2007, background): “Index Terms color image denoising, adaptive grouping, blockmatching, shrinkage.”
Kernel-predicting convolutional networks for denoising Monte Carlo renderings (2017, neutral): “Indeed, more powerful image-space lters, such as BM3D [Dabov et al. 2006], have seen less use for MC denoising with some notable exceptions [Kalantari and Sen 2013].”
Applications, promises, and pitfalls of deep learning for fluorescence image reconstruction (2018, methodology): “CARE maintains image quality while decreasing light exposure 60-fold and outperforms classical denoising such as nonlocal means 30 or block-matching and three-dimensional (3D) filtering 31 .”
SAR Image Despeckling Using a Convolutional Neural Network (2017, methodology): “Various methods have been developed in the literature to suppress speckle including multilook processing [3], [4], filtering methods [5]–[7], wavelet-based despecking methods [8]–[11], block-matching 3D (BM3D) algorithm [12] and total variation (TV) methods [13].”
Deep Retinex Decomposition for Low-Light Enhancement (2018, methodology): “We collect 270 low-light images from public MEF [18], NPE [23], LIME [9], DICM [13], VV 1 , and Fusion [3] dataset, transform the images into YCbCr channel and calculate the histogram of Y channel.”
Neural Nearest Neighbors Networks (2018, methodology): “These include the classical BM3D [9], which uses a hand-crafted denoising pipeline, and the state-of-the-art trainable non-local models NLNet [23] and UNLNet [24], both learning to process non-locally aggregated patches.”