PCA Gaussianization for image processing

Laparra, Valero; Camps‐Valls, Gustau; Malo, Jesús

doi:10.1109/icip.2009.5413808

Cited by 9 publications

(10 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…120000 pairs of coefficients were used in each MI estimation. Two kinds of MI estimators were used: (1) direct computation of MI, which involves 2D histogram estimation [43], and (2) estimation of MI by PCA-based Gaussianization (GPCA) [44], which only involves univariate histogram estimations. Table 2 shows the MI results (in bits) for pairs of coefficients in the wavelet and the divisive normalized domains.…”

Section: D Statistical Effect Of the Divisive Normalizationmentioning

confidence: 99%

Divisive normalization image quality metric revisited

2010

Self Cite

View full text Add to dashboard Cite

Structural similarity metrics and information-theory-based metrics have been proposed as completely different alternatives to the traditional metrics based on error visibility and human vision models. Three basic criticisms were raised against the traditional error visibility approach: (1) it is based on near-threshold performance, (2) its geometric meaning may be limited, and (3) stationary pooling strategies may not be statistically justified. These criticisms and the good performance of structural and information-theory-based metrics have popularized the idea of their superiority over the error visibility approach. In this work we experimentally or analytically show that the above criticisms do not apply to error visibility metrics that use a general enough divisive normalization masking model. Therefore, the traditional divisive normalization metric 1 is not intrinsically inferior to the newer approaches. In fact, experiments on a number of databases including a wide range of distortions show that divisive normalization is fairly competitive with the newer approaches, robust, and easy to interpret in linear terms. These results suggest that, despite the criticisms of the traditional error visibility approach, divisive normalization masking models should be considered in the image quality discussion.

show abstract

Section: D Statistical Effect Of the Divisive Normalizationmentioning

confidence: 99%

Divisive normalization image quality metric revisited

2010

Self Cite

View full text Add to dashboard Cite

show abstract

“…In the utilized thresholding function, which is based on maximu m a posteriori (MAP) estimate, the modified version of dominant coefficients was estimated by optimal linear interpolation between each coefficient and the mean value of the corresponding sub band [31]. In this wo rk, authors proposed a fast alternative to iterative Gaussianization methods that makes it suitable for image processing while ensuring its theoretical convergence [32]. Author performed a co mparison between two source signal extract ion algorithms, namely the Wavelet De-noising (WD) by Soft Thresholding and Independent Co mponent Analysis (ICA) on a simu lated functional optical imaging data.…”

Section: Literature Reviewmentioning

confidence: 99%

Spatial and Transform Domain Filtering Method for Image De-noising: A Review

Roy¹

2013

IJMECS

View full text Add to dashboard Cite

Present investigation reveals the quantum of work carried in the filtering methods for image denoising. An image is often gets corrupted by various noises that are visible or invisible while being gathered, coded, acquired and transmitted. Noise influences various process parameters that may cause a quality problem for further image processing. De-noising of natural images is appears to be very simp le however when considered under practical situations becomes complex. It has been cited by various author that parameter such as type and quantum of noise, image etc. through single algorith m or approach becomes cumbersome when results are optimized. In order to improve the quality of an image noise must be removed when the image is pre-p rocessed and the important signal features like edge details should be retained as much as possible. The search on efficient image denoising methods is still a valid challenge at the crossing of functional analysis and statistics. This paper reviews significant de-noising methods (spatial and transform domain method) and their salient features and applications. One filter in each category has been taken in consideration to understand the characteristics of both spatial and transform do main filters.

show abstract

“…Instead of estimating the biases of the clusters jointly 3 we follow a greedy approach. First, estimate the bias b j of each cluster C j independently, as the minimizer of the norm of the mean of the Gaussianized data 4 ,…”

Section: Radial Gaussianization After Bias Compensationmentioning

confidence: 99%

“…3 Joint optimization is a slow process when the optimization function is not given in closed form or is not differentiable. 4 We also experimented with an alternative where the mean of each cluster C j was used as an estimate of its bias. Our experiments showed that this did not improve RG due to the fact that the clusters were not estimated perfectly.…”

Section: Radial Gaussianization After Bias Compensationmentioning

confidence: 99%

See 1 more Smart Citation

Radial Gaussianization with cluster-specific bias compensation

Huang

Karakos

2011

2011 IEEE Statistical Signal Processing Workshop (SSP)

View full text Add to dashboard Cite

In recent work, Lyu and Simoncelli [1] introduced radial Gaussianization (RG) as a very efficient procedure for transforming n-dimensional random vectors into Gaussian vectors with independent and identically distributed (i.i.d.) components. This entails transforming the norms of the data so that they become chi-distributed with n degrees of freedom. A necessary requirement is that the original data are generated by an isotropic distribution, that is, their probability density function (pdf) is constant over surfaces of n-dimensional spheres (or, more general, n-dimensional ellipsoids). The case of biases in the data, which is of great practical interest, is studied here; as we demonstrate with experiments, there are situations in which even very small amounts of bias can cause RG to fail. This becomes evident especially when the data form clusters in low-dimensional manifolds. To address this shortcoming, we propose a two-step approach which entails (i) first discovering clusters in the data and removing the bias from each, and (ii) performing RG on the bias-compensated data. In experiments with synthetic data, the proposed bias compensation procedure results in significantly better Gaussianization than the non-compensated RG method.

show abstract

PCA Gaussianization for image processing

Cited by 9 publications

References 12 publications

Divisive normalization image quality metric revisited

Divisive normalization image quality metric revisited

Spatial and Transform Domain Filtering Method for Image De-noising: A Review

Radial Gaussianization with cluster-specific bias compensation

Contact Info

Product

Resources

About