In this paper we propose a combined method for removing impulse noise. In the first phase, we use an efficient detector, called the statistics of Ordered Difference Detector (SODD) to identify pixels which are likely ...
详细信息
ISBN:
(纸本)9783540728221
In this paper we propose a combined method for removing impulse noise. In the first phase, we use an efficient detector, called the statistics of Ordered Difference Detector (SODD) to identify pixels which are likely to be corrupted by impulse noise. The proposed SODD can yield very high noise detection accuracy at high noise density. This noise detection phase is crucial for the following noise removal. In the second phase, only these noise candidates are restored using the variational method. Edges and noise free pixels of images filtered by our combined method are preserved. Simulation results indicate that the proposed method is significantly better than those using other impulse noise reduction filters.
We present a novel geometric approach for solving the stereo problem for an arbitrary number of images (greater than or equal to 2). It is based upon the definition of a variational principle that must be satisfied by...
详细信息
ISBN:
(纸本)3540631674
We present a novel geometric approach for solving the stereo problem for an arbitrary number of images (greater than or equal to 2). It is based upon the definition of a variational principle that must be satisfied by the surfaces of the objects in the scene and their images. The Euler-Lagrange equations which are deduced from the variational principle provide a set of PDE's which are used to deform an initial set of surfaces which then move towards the objects to be detected. The level set implementation of these PDE's potentially provides an efficient and robust way of achieving the surface evolution and to deal automatically with changes in the surface topology during the deformation, i.e. to deal with multiple objects. Results of a two dimensional implementation of our theory are presented on synthetic and real images.
This is a theoretical study on the minimizers of cost-functions composed of an l(2) data-fidelity term and a possibly nonsmooth or non-convex regularization term acting on the differences or the discrete gradients of ...
详细信息
ISBN:
(纸本)9783540728221
This is a theoretical study on the minimizers of cost-functions composed of an l(2) data-fidelity term and a possibly nonsmooth or non-convex regularization term acting on the differences or the discrete gradients of the image or the signal to restore. More precisely, we derive general nonasymptotic analytical bounds characterizing the local and the global minimizers of these cost-functions. We provide several bounds relevant to the observation model. For edge-preserving regularization, we exhibit a tight bound on the l(infinity) norm of the residual (the error) that is independent of the data, even if its l(2) norm is being minimized. Then we focus on the smoothing incurred by the (local) minimizers in terms of the differences or the discrete gradient of the restored image (or signal).
Inverse scalespacemethods are derived as asymptotic limits of iterative regularization methods. They have proven to be efficient methods for denoising of gray valued images and for the evaluation of unbounded operat...
详细信息
ISBN:
(纸本)9783540728221
Inverse scalespacemethods are derived as asymptotic limits of iterative regularization methods. They have proven to be efficient methods for denoising of gray valued images and for the evaluation of unbounded operators. In the beginning, inverse scalespacemethods have been derived from iterative regularization methods with squared Hilbert norm regularization terms, and later this concept was generalized to Bregman distance regularization (replacing the squared regularization norms);therefore allowing for instance to consider iterative total variation regularization. We have proven recently existence of a solution of the associated inverse total variation flow equation. In this paper we generalize these results and prove existence of solutions of inverse flow equations derived from iterative regularization with general convex regularization functionals. We present some applications to filtering of color data and for the stable evaluation of the diZenzo edge detector.
A geometric model is proposed for an artificial foveal vision system, and its plausibility in the context of biological vision is explored. The model is based on an isotropic, scale invariant two-form that describes t...
详细信息
ISBN:
(纸本)9783540728221
A geometric model is proposed for an artificial foveal vision system, and its plausibility in the context of biological vision is explored. The model is based on an isotropic, scale invariant two-form that describes the spatial layout of receptive fields in the the visual sensorium (in the biological context roughly corresponding to retina, LGN, and V1). It overcomes the limitation of the familiar log-polar model by handling its singularity in a graceful way. The log-polar singularity arises as a result of ignoring the physical resolution limitation inherent in any real (artificial or biological) visual system. The incorporation of such a limitation requires the introduction of a physical constant, measuring the radius of the geometric foveola (a central region characterized by maximal resolving power). The proposed model admits a description in singularity-free canonical coordinates that generalize the well-established log-polar coordinates, and that reduce to these in the asymptotic case of negligibly sized geometric foveola (or, equivalently, at peripheral locations in the visual field). Biological plausibility of the model is demonstrated by comparison with known facts on human vision.
Curve evolution forms the basis of active contour algorithms used for image segmentation. In many applications the curve under evolution needs to be restricted to the shape space given by some example shapes, or some ...
详细信息
ISBN:
(纸本)9783540728221
Curve evolution forms the basis of active contour algorithms used for image segmentation. In many applications the curve under evolution needs to be restricted to the shape space given by some example shapes, or some linear space given by a set of basis vectors. Also, when a curve evolution is carried out on a computer, the evolution is approximated by some suitable discretization. Here too, the evolution is implicitly carried out in some subspace and not in the space of all curves. Hence it is important to study curve evolution in subspace of all curves. We look at a formulation that describes curve evolution restricted to subspaces. We give numerical methods and examples of a formulation for curvature flow for curves restricted to the B-spline subspace.
A variational formulation of an image analysis problem has the nice feature that it is often easier to predict the effect of minimizing a certain energy functional than to interpret the corresponding Euler-Lagrange eq...
详细信息
ISBN:
(纸本)9783540728221
A variational formulation of an image analysis problem has the nice feature that it is often easier to predict the effect of minimizing a certain energy functional than to interpret the corresponding Euler-Lagrange equations. For example, the equations of motion for an active contour usually contains a mean curvature term, which we know will regularizes the contour because mean curvature is the first variation of curve length, and shorter curves are typically smoother than longer ones. In some applications it may be worth considering Gaussian curvature as a regularizing term instead of mean curvature. The present paper provides a variational principle for this: We show that Gaussian curvature of a regular surface in three-dimensional Euclidean space is the first variation of an energy functional defined on the surface. Some properties of the corresponding motion by Gaussian curvature are pointed out, and a simple example is given, where minimization of this functional yields a nontrivial solution.
Although variational models offer many advantages in image analysis, their successful application to real-world problems is documented only for some specific areas such as medical imaging. In this paper we show how we...
详细信息
ISBN:
(纸本)9783540728221
Although variational models offer many advantages in image analysis, their successful application to real-world problems is documented only for some specific areas such as medical imaging. In this paper we show how well-adapted variational ideas can solve the problem of hairstyle simulation in a fully automatic way: A customer in a hairdresser's shop selects a new hairstyle from a database, and this hairstyle is automatically registered to a digital image of the customer's face. Interestingly already a carefully modified optic flow method of Horn and Schunck turns out to be ideal for this application. These modifications include an extension to colour sequences, an incorporation of warping ideas in order to allow large deformation rates, and the inclusion of shape information that is characteristic for human faces. Employing classical numerical ideas such as finite differences and SOR iterations offers sufficient performance for real-life applications. In a number of experiments we demonstrate that our variational approach is capable of solving the hairstyle simulation problem with high quality in a fully practical setting.
In this paper, we advance the state of the art in variational image segmentation through the fusion of bottom-up segmentation and top-down classification of object behavior over an image sequence. Such an approach is ...
详细信息
ISBN:
(纸本)9783540728221
In this paper, we advance the state of the art in variational image segmentation through the fusion of bottom-up segmentation and top-down classification of object behavior over an image sequence. Such an approach is beneficial for both tasks and is carried out through a joint optimization, which enables the two tasks to cooperate, such that knowledge relevant to each can aid in the resolution of the other, thereby enhancing the final result. In particular, classification offers dynamic probabilistic priors to guide segmentation, while segmentation supplies its results to classification, ensuring that they are consistent with prior knowledge. The prior models are learned from training data and updated dynamically, based on segmentations of earlier images in the sequence. We demonstrate the potential of our approach in a hand gesture recognition application, where the combined use of segmentation and classification improves robustness in the presence of occlusion and background complexity.
In this paper, we present an adaptive variational segmentation algorithm of spectral / texture regions in satellite images using level set. Satellite images contain both textured and non-textured regions, so for each ...
详细信息
ISBN:
(纸本)9783540728221
In this paper, we present an adaptive variational segmentation algorithm of spectral / texture regions in satellite images using level set. Satellite images contain both textured and non-textured regions, so for each region spectral and texture cues are integrated according to their discrimination power. Motivated by Fisher-Rao linear discriminant analysis;two region weights are defined to code respectively the relevance of spectral and texture cues. Therefore;regions with or without texture are processed in an unified framework. The obtained segmentation criterion is minimized via curves evolution within an explicit correspondence between the interiors of evolving curves and regions in the segmentation. The shape derivation principle is used to derive the system of coupled evolution equations in such a way that we consider the region weights and the statistical parameters variability. Experimental results on both natural and satellite images are shown.
暂无评论