Jaime Gutierrez

Perceptual feedback in multigrid motion estimation using an improved DCT quantization

In this paper, a multigrid motion compensation video coder based on the current human visual system (HVS) contrast discrimination models is proposed. A novel procedure for the encoding of the prediction errors has been used. This procedure restricts the maximum perceptual distortion in each transform coefficient. This subjective redundancy removal procedure includes the amplitude nonlinearities and some temporal features of human perception. A perceptually weighted control of the adaptive motion estimation algorithm has also been derived from this model. Perceptual feedback in motion estimation ensures a perceptual balance between the motion estimation effort and the redundancy removal process. The results show that this feedback induces a scale-dependent refinement strategy that gives rise to more robust and meaningful motion estimation, which may facilitate higher level sequence interpretation. Perceptually meaningful distortion measures and the reconstructed frames show the subjective improvements of the proposed scheme versus an H.263 scheme with unweighted motion estimation and MPEG-like quantization.

Linear transform for simultaneous diagonalization of covariance and perceptual metric matrix in image coding

Two types ofredundancies are contained in images: statistical redundancy and psychovisual redundancy. Image representation techniques for image coding should remove both redundancies in order to obtain good results. In order to establish an appropriate representation, the standard approach to transform coding only considers the statistical redundancy, whereas the psychovisual factors are introduced after the selection ofthe representation as a simple scalar weighting in the transform domain. In this work, we take into account the psychovisual factors in the de8nition of the representation together with the statistical factors, by means of the perceptual metric and the covariance matrix, respectively. In general the ellipsoids described by these matrices are not aligned. Therefore, the optimal basis for image representation should simultaneously diagonalize both matrices. This approach to the basis selection problem has several advantages in the particular application ofimage coding. As the transform domain is Euclidean (by de8nition), the quantizer design is highly simpli8ed and at the same time, the use ofscalar quantizers is truly justi8ed. The proposed representation is compared to covariance-based representations such as the DCT and the KLT or PCA using standard JPEG-like and Max-Lloyd quantizers. ? 2003 Pattern Recognition Society. Published by Elsevier Science Ltd. All rights reserved.

Regularization operators for natural images based on nonlinear perception models

Image restoration requires some a priori knowledge of the solution. Some of the conventional regularization techniques are based on the estimation of the power spectrum density. Simple statistical models for spectral estimation just take into account second-order relations between the pixels of the image. However, natural images exhibit additional features, such as particular relationships between local Fourier or wavelet transform coefficients. Biological visual systems have evolved to capture these relations. We propose the use of this biological behavior to build regularization operators as an alternative to simple statistical models. The results suggest that if the penalty operator takes these additional features in natural images into account, it will be more robust and the choice of the regularization parameter is less critical.

Perceptual adaptive insensitivity for support vector machine image coding

Support vector machine (SVM) learning has been recently proposed for image compression in the frequency domain using a constant /spl epsiv/-insensitivity zone by Robinson and Kecman. However, according to the statistical properties of natural images and the properties of human perception, a constant insensitivity makes sense in the spatial domain but it is certainly not a good option in a frequency domain. In fact, in their approach, they made a fixed low-pass assumption as the number of discrete cosine transform (DCT) coefficients to be used in the training was limited. This paper extends the work of Robinson and Kecman by proposing the use of adaptive insensitivity SVMs for image coding using an appropriate distortion criterion , based on a simple visual cortex model. Training the SVM by using an accurate perception model avoids any a priori assumption and improves the rate-distortion performance of the original approach.

Image Denoising with Kernels Based on Natural Image Relations

A successful class of image denoising methods is based on Bayesian approaches working in wavelet representations. The performance of these methods improves when relations among the local frequency coefficients are explicitly included. However, in these techniques, analytical estimates can be obtained only for particular combinations of analytical models of signal and noise, thus precluding its straightforward extension to deal with other arbitrary noise sources. In this paper, we propose an alternative non-explicit way to take into account the relations among natural image wavelet coefficients for denoising: we use support vector regression (SVR) in the wavelet domain to enforce these relations in the estimated signal. Since relations among the coefficients are specific to the signal, the regularization property of SVR is exploited to remove the noise, which does not share this feature. The specific signal relations are encoded in an anisotropic kernel obtained from mutual information measures computed on a representative image database. In the proposed scheme, training considers minimizing the Kullback-Leibler divergence (KLD) between the estimated and actual probability functions (or histograms) of signal and noise in order to enforce similarity up to the higher (computationally estimable) order. Due to its non-parametric nature, the method can eventually cope with different noise sources without the need of an explicit re-formulation, as it is strictly necessary under parametric Bayesian formalisms. Results under several noise levels and noise sources show that: (1) the proposed method outperforms conventional wavelet methods that assume coefficient independence, (2) it is similar to state-of-the-art methods that do explicitly include these relations when the noise source is Gaussian, and (3) it gives better numerical and visual performance when more complex, realistic noise sources are considered. Therefore, the proposed machine learning approach can be seen as a more flexible (model-free) alternative to the explicit description of wavelet coefficient relations for image denoising.

An active contour model for the automatic detection of the fovea in fluorescein angiographies

Fovea segmentation in fluorescein angiographies is a fundamental first task in any study of ocular diseases. The importance of fovea detection is due to the fact that the nearer the centre of the fovea a lesion is, the graver this lesion is. The proposed method is based on B-snakes and uses a greedy algorithm to minimise an appropriate energy which accurately leads to a convenient characterisation of the boundary of the foveal zone. The first initialisation step, which consists of finding the most appropriate local minimum along with a procedure to construct an initial contour involving a region growing algorithm, leads to a convenient and robust initialisation of the proposed active contour model.

Recovering wavelet relations using SVM for image denoising

Here we propose an alternative non-explicit way to take into account the relations among wavelet coefficients in natural images for denoising: we use support vector machines (SVM) to learn these relations. Since relations among the coefficients are specific to the signal, SVM regularization removes the noise, which does not share this property. Moreover, due to its non-parametric nature, the method can eventually cope with different noise sources. The results show that: (1) the proposed non-parametric method outperforms conventional methods that assume coefficient independence, and (2) its performance is similar to state-of-the-art parametric methods that do explicitly include these relations. Therefore, the proposed machine learning approach can be seen as a more flexible (model-free) alternative to the explicit description of wavelet coefficient relations in Bayesian approaches.

Set Descriptors for Visual Evaluation of Human Corneal Endothelia

Images of corneal endothelium obtained from specular microscopy are of great importance in the evaluation of the corneal endothelium status. Several commercial tools provide some numerical descriptors to characterize these images in terms of cell density, hexagonality, and some descriptive statistics of the cell areas. However, it is a too simple analysis that only detects severe abnormal endothelia with many irregular and large cells. Detection of subtle abnormalities needs a more refined analysis. This paper proposes a shape-size descriptor based on some modified versions of the geometric covariogram. This descriptor is presented as a valid alternative to the classical analysis that provides a reliable, visual, and easy evaluation of the corneal endothelium. Control images of normal endothelia can be used to classify a given image as normal or abnormal with respect to these controls by means of a graphical test.

Perceptual regularization functionals for natural image restoration

Regularization constraints are necessary in inverse problems such as image restoration, optical flow computation or shape from shading to avoid the singularities in the solution. Conventional regularization techniques are based on some a priori knowledge of the solution: usually, the solution is assumed to be smooth according to simple statistical image or motion models. Using the fact that human visual perception is adapted to the statistics of natural images and sequences, the class of regularization functionals proposed in this work are not based on an image model but on a model of the human visual system. In particular, the current nonlinear model of early human visual processing is used to obtain locally adaptive regularization functionals for image restoration without any a priori assumption on the image or the noise. The results show that these functionals constitute a valid alternative to those based on the local autocorrelation of the image.

A Color Contrast Definition for Perceptually Based Color Image Coding

The non-linear nature of the human visual response to achromatic contrast is a key element to improve the performance in achromatic image coding. Expressing transform coefficients in the appropriate contrast units is relevant when some particular non-linear processing hasto be applied. In the achromatic case, the use of non-linear psychophysical models is straightforward since achromatic contrast computation from image transform coefficients is quite simple. However, using equivalent color masking models in transform coding is not easy since psychophysical results are expressed in color contrast units which are non-trivially related to the transform coefficients in opponent color spaces. In this patent we describe a general procedure to define color contrast for any spatial basis functions (such as block-DCT or wavelets) with any chromatic modulation. The proposed definition is based on (1) simple psychophysics to define purely chromatic basis functions, and (2) statistical analysis of the chromatic content of natural images to define the maximum chromatic modulation. The proposed color contrast definition allows for a straightforward extension of the well known non-linear achromatic masking models to the chromatic case for color image coding. In this work, the use of the proposed color contrast definition is illustrated by a particular non-linear color image coding scheme based on blockDCT, non-linear perceptual response transforms in YUV color channels, and non-linear machine learning response selection. This non-linear scheme is compared to the equivalent linear (JPEG-like) scheme, where color contrast definition is not relevant due to its linear nature.