Noise estimation for video processing based on spatio-temporal gradients

Zlokolica, Vladimir; Pižurica, Aleksandra; Philips, Wilfried

doi:10.1109/lsp.2006.870481

Cited by 34 publications

(17 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For gray level images, the MAD framework [19] uses the deviation from a smooth image model to estimate the noise. Specifically, two state-of-the-art methods [2,10] take a similar approach.…”

Section: Estimating Noise Standard Deviation In Imagesmentioning

confidence: 99%

Scale invariance and noise in natural images

Zoran

Weiss

2009

2009 IEEE 12th International Conference on Computer Vision

212

135

View full text Add to dashboard Cite

Natural images are known to have scale invariant statistics. While some eariler studies have reported the kurtosis of marginal bandpass filter response distributions to be constant throughout scales, other studies have reported that the kurtosis values are lower for high frequency filters than for lower frequency ones. In this work we propose a resolution for this discrepancy and suggest that this change in kurtosis values is due to noise present in the image. We suggest that this effect is consistent with a clean, natural image corrupted by white noise. We propose a model for this effect, and use it to estimate noise standard deviation in corrupted natural images. In particular, our results suggest that classical benchmark images used in low-level vision are actually noisy and can be cleaned up. Our results on noise estimation on two sets of 50 and a 100 natural images are significantly better than the state-of-the-art.

show abstract

Section: Estimating Noise Standard Deviation In Imagesmentioning

confidence: 99%

Scale invariance and noise in natural images

Zoran

Weiss

2009

2009 IEEE 12th International Conference on Computer Vision

212

135

View full text Add to dashboard Cite

show abstract

“…Here, it was possible to prove the accuracy of our algorithm using original data. The other problem is to have an accurate model of noise because the majority of algorithms that compute the noise are based on a model of the distribution of this noise (Liu et al 2006;Zlokolica et al 2006). Without an a priori on this distribution, we decided, in our case, to use an algorithm to find the best lagrangian multiplier.…”

Section: Estimation Of Kmentioning

confidence: 99%

High Quality Real-Time Video with Scanning Electron Microscope Using Total Variation Algorithm on a Graphics Processing Unit

Ouarti

Sauvet

Régnier

2012

International Journal of Optomechatronics

View full text Add to dashboard Cite

The scanning electron microscope (SEM) is usually dedicated to taking a picture of micro-nanoscopic objects. In the present study, we wondered whether a SEM can be converted as a real-time video display. To this end, we designed a new methodology. We use the slow mode of the SEM to acquire a high quality reference image that can then be used to estimate the optimal parameters that regularize the signal for a given method. Here, we employ Total Variation, a method which minimizes the noise and regularizes the image. An optimal lagrangian multiplier can be computed that regularizes the image efficiently. We showed that a limited number of iterations for Total Variation algorithm can lead to an acceptable quality of regularization. This algorithm is parallel and deployed on a Graphics Processing Unit to obtain a real-time high quality video with a SEM. It opens the possibility of a real-time interaction at micro-nanoscales.

show abstract

“…The proposed technique shows [14] efficient and accurate wavelet based noise estimation method for white Gaussian noise in video sequences. Instead of motion estimation step, the surfacelet transform provides motion selective subband decomposition for video signals and it is low redundancy [15].…”

Section: Video Denoising -Related Workmentioning

confidence: 99%

Noise Reduction in Video Sequences The State of Art and the Technique for Motion Detection

Reeja¹,

Kavya²

2012

IJCA

View full text Add to dashboard Cite

This paper provides a detailed state of the art of different video denoising techniques. Most of the video denoising algorithms are done through the motion detection technique. The main goal is to give a survey of various noise reduction techniques for video. Object detection is the first level of video denoising. The first level can be achieved through Motion Detection. This paper explained about the motion estimation and compensation techniques. The different video denoising techniques, motion detection techniques and the noises used are shown through the taxonomy. General TermsImage processing, motion detection KeywordsObject tracking, frame differencing, image and video denoising.

show abstract

Noise estimation for video processing based on spatio-temporal gradients

Cited by 34 publications

References 11 publications

Scale invariance and noise in natural images

Scale invariance and noise in natural images

High Quality Real-Time Video with Scanning Electron Microscope Using Total Variation Algorithm on a Graphics Processing Unit

Noise Reduction in Video Sequences The State of Art and the Technique for Motion Detection

Contact Info

Product

Resources

About