Automatic Parameter Selection for Denoising Algorithms Using a No-Reference Measure of Image Content

By Xiang Zhu, Peyman Milanfar. IEEE Transactions on Image Processing, 2010

396 citations recorded by the source corpus.

Subfields: Computer Vision and Pattern Recognition, Media Technology

View Paper: https://www.semanticscholar.org/paper/d81bc1b04d17af15cdc21b259eb099d6d8ca1502

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

No-Reference Image Quality Assessment in the Spatial Domain (2012, methodology): “We note that our training approach resembles that of [49].”
From Denoising to Compressed Sensing (2014, methodology): “Toward this goal, we can employ different strategies that have been proposed in the literature for setting the parameters of denoising algorithms [63]–[65].”
Single-Image Noise Level Estimation for Blind Denoising (2013, methodology): “For the second category, Zhu and Milanfar [11] proposed a non-reference metric Q based on image gradient statistics.”
NIMA: Neural Image Assessment (2017, result): “Application of our no-reference quality metric as a prior in image denoising is similar to the work of Zhu et al. [40], [41].”
No-Reference Image Sharpness Assessment in Autoregressive Parameter Space (2015, background): “Furthermore, this type of approaches are of many valuable applications in image processing, such as automatic contrast enhancement [9], [10], super-resolution [11] and denoising [12].”
Denoising of 3D magnetic resonance images by using higher-order singular value decomposition (2015, background): “Stein’s unbiased risk estimate (Stein, 1981) or the no-reference metric Q (Zhu and Milanfar, 2010) can serve as quantitative metrics for automatic parameter tuning when no ‘‘ground-truth’’ reference is available, but will require a number of iterations.”
dipIQ: Blind Image Quality Assessment by Learning-to-Rank Discriminable Image Pairs (2017, background): “The singular value decomposition of local image gradient matrices may provide a quantitative measure of image content [42].”
A no-reference metric for evaluating the quality of motion deblurring (2013, methodology): “In previous work image quality metrics have been applied to automatically select good parameter settings for image processing algorithms, such as image denoising [Zhu and Milanfar 2010; Mittal et al. 2012a].”
From Heuristic Optimization to Dictionary Learning: A Review and Comprehensive Comparison of Image Denoising Algorithms (2014, methodology): “4) Overcomplete image models are better (KSPR is better than MSKR at high noise levels).”
Hyperspectral Image Denoising Employing a Spectral–Spatial Adaptive Total Variation Model (2012, methodology): “A recent blind image content measurement index called the Q-metric [48] is used to evaluate the spatial information preservation.”