Superpixel-based image noise variance estimation with local statistical assessment

Wu, Cheng-Ho; Chang, Herng-Hua

doi:10.1186/s13640-015-0093-2

Cited by 21 publications

(9 citation statements)

References 33 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Noise-level estimation from a single image is an illposed problem because it is impossible to completely separate textures from noise, and many methods have been developed to tackle this problem [9,10,15,25]. Some methods succeeded in estimating noise level by using patches sampled from homogeneous areas [2,20,21,22]. These methods are based on the assumption that there is a suffi-cient amount of flat areas in the input image, but this assumption does not necessarily hold in natural images with rich textures.…”

Section: Related Workmentioning

confidence: 99%

Noise-Level Estimation from Single Color Image Using Correlations Between Textures in RGB Channels

Nakamura¹,

Kobayashi²

2019

Preprint

View full text Add to dashboard Cite

We propose a simple method for estimating noise level from a single color image. In most image-denoising algorithms, an accurate noise-level estimate results in good denoising performance; however, it is difficult to estimate noise level from a single image because it is an ill-posed problem. We tackle this problem by using prior knowledge that textures are highly correlated between RGB channels and noise is uncorrelated to other signals. We also extended our method for RAW images because they are available in almost all digital cameras and often used in practical situations. Experiments show the high noise-estimation performance of our method in synthetic noisy images. We also applied our method to natural images including RAW images and achieved better noise-estimation performance than conventional methods.

show abstract

Section: Related Workmentioning

confidence: 99%

Noise-Level Estimation from Single Color Image Using Correlations Between Textures in RGB Channels

Nakamura¹,

Kobayashi²

2019

Preprint

View full text Add to dashboard Cite

show abstract

“…In contrast, we propose a simple and efficient method to estimate σ 2 in an observed point cloud. Our approach is analogous to [53] that estimates the noise variance of natural images. Assuming an iid noise model, [53] proposes to: i) identify flat patches (each patch has approximately a constant pixel value) from the given image, and ii) estimate the noise variance locally in each flat patch.…”

Section: A Weight Parameter and Noise Variancementioning

confidence: 99%

“…Our approach is analogous to [53] that estimates the noise variance of natural images. Assuming an iid noise model, [53] proposes to: i) identify flat patches (each patch has approximately a constant pixel value) from the given image, and ii) estimate the noise variance locally in each flat patch. Analogously, we first identify approximately flat 3D patches in a given noisy point cloud, and then estimate the noise variance from each patch.…”

Section: A Weight Parameter and Noise Variancementioning

confidence: 99%

Local 3D Point Cloud Denoising via Bipartite Graph Approximation & Total Variation

Dinesh

Cheung

Bajić

et al. 2018

2018 IEEE 20th International Workshop on Multimedia Signal Processing (MMSP)

View full text Add to dashboard Cite

Point cloud is a collection of 3D coordinates that are discrete geometric samples of an object's 2D surfaces. Imperfection in the acquisition process means that point clouds are often corrupted with noise. Building on recent advances in graph signal processing, we design local algorithms for 3D point cloud denoising. Specifically, we design a reweighted graph Laplacian regularizer (RGLR) for surface normals and demonstrate its merits in rotation invariance, promotion of piecewise smoothness, and ease of optimization. Using RGLR as a signal prior, we formulate an optimization problem with a general p-norm fidelity term that can explicitly model two types of independent noise: small but non-sparse noise (using 2 fidelity term) and large but sparser noise (using 1 fidelity term).To establish a linear relationship between normals and 3D point coordinates, we first perform bipartite graph approximation to divide the point cloud into two disjoint node sets (red and blue). We then optimize the red and blue nodes' coordinates alternately. For 2-norm fidelity term, we iteratively solve an unconstrained quadratic programming (QP) problem, efficiently computed using conjugate gradient with a bounded condition number to ensure numerical stability. For 1-norm fidelity term, we iteratively minimize an 1-2 cost function sing accelerated proximal gradient (APG), where a good step size is chosen via Lipschitz continuity analysis. Finally, we propose simple mean and median filters for flat patches of a given point cloud to estimate the noise variance given the noise type, which in turn is used to compute a weight parameter trading off the fidelity term and signal prior in the problem formulation. Extensive experiments show state-of-the-art denoising performance among local methods using our proposed algorithms.

show abstract

“…Based on the superpixel properties around the region edges, Li et al [19] proposed a method of superpixel-guided non-local means for natural image noise reduction. Superpixel models have also been introduced to estimate the image noise level [20,21]. As the superpixels can be used to segment the images effectively, using the local statistics of superpixels to recover the images may be another feasible way for speckle noise reduction from ultrasound images.…”

Section: Introductionmentioning

confidence: 99%

A New Ultrasound Speckle Reduction Algorithm Based on Superpixel Segmentation and Detail Compensation

et al. 2019

View full text Add to dashboard Cite

Speckle is a kind of noise commonly found in ultrasound images (UIs). Although traditional local operation-based methods, such as bilateral filtering, perform well in de-noising normal natural images with suitable parameters, these methods may break local correlations and, hence, their performance will be highly degraded when applied to UIs with high levels of speckle noise. In this work, we propose a new method, based on superpixel segmentation and detail compensation, to reduce UI speckle noise. In particular, considering that superpixel segmentation has the advantage of adhering accurately to the boundaries of objects or local structures, we propose a superpixel version of bilateral filtering to better protect the local structure during de-noising. Additionally, a human visual system (HVS)-inspired strategy for spatial compensation is introduced, in order to recover sophisticated edges as much as possible while weakening the high-frequency noise. Experiments on synthetic images and real UIs of different organs show that, compared to other methods, the proposed strategy can reduce ultrasound speckle noise more effectively.

show abstract

Superpixel-based image noise variance estimation with local statistical assessment

Cited by 21 publications

References 33 publications

Noise-Level Estimation from Single Color Image Using Correlations Between Textures in RGB Channels

Noise-Level Estimation from Single Color Image Using Correlations Between Textures in RGB Channels

Local 3D Point Cloud Denoising via Bipartite Graph Approximation & Total Variation

A New Ultrasound Speckle Reduction Algorithm Based on Superpixel Segmentation and Detail Compensation

Contact Info

Product

Resources

About