Results 1 
7 of
7
Mean shift: A robust approach toward feature space analysis
 In PAMI
, 2002
"... A general nonparametric technique is proposed for the analysis of a complex multimodal feature space and to delineate arbitrarily shaped clusters in it. The basic computational module of the technique is an old pattern recognition procedure, the mean shift. We prove for discrete data the convergence ..."
Abstract

Cited by 1461 (34 self)
 Add to MetaCart
A general nonparametric technique is proposed for the analysis of a complex multimodal feature space and to delineate arbitrarily shaped clusters in it. The basic computational module of the technique is an old pattern recognition procedure, the mean shift. We prove for discrete data the convergence of a recursive mean shift procedure to the nearest stationary point of the underlying density function and thus its utility in detecting the modes of the density. The equivalence of the mean shift procedure to the Nadaraya–Watson estimator from kernel regression and the robust Mestimators of location is also established. Algorithms for two lowlevel vision tasks, discontinuity preserving smoothing and image segmentation are described as applications. In these algorithms the only user set parameter is the resolution of the analysis, and either gray level or color images are accepted as input. Extensive experimental results illustrate their excellent performance.
NavierStokes, fluid dynamics, and image and video inpainting
 Proc. IEEE Computer Vision and Pattern Recognition (CVPR
, 2001
"... Image inpainting involves filling in part of an image or video using information from the surrounding area. Applications include the restoration of damaged photographs and movies and the removal of selected objects. In this paper, we introduce a class of automated methods for digital inpainting. The ..."
Abstract

Cited by 112 (16 self)
 Add to MetaCart
Image inpainting involves filling in part of an image or video using information from the surrounding area. Applications include the restoration of damaged photographs and movies and the removal of selected objects. In this paper, we introduce a class of automated methods for digital inpainting. The approach uses ideas from classical fluid dynamics to propagate isophote lines continuously from the exterior into the region to be inpainted. The main idea is to think of the image intensity as a ‘stream function ’ for a twodimensional incompressible flow. The Laplacian of the image intensity plays the role of the vorticity of the fluid; it is transported into the region to be inpainted by a vector field defined by the stream function. The resulting algorithm is designed to continue isophotes while matching gradient vectors at the boundary of the inpainting region. The method is directly based on the NavierStokes equations for fluid dynamics, which has the immediate advantage of welldeveloped theoretical and numerical results. This is a new approach for introducing ideas from computational fluid dynamics into problems in computer vision and image analysis.
Global Computational Algebraic Topology Approach for Diffusion
, 2004
"... One physical process involved in many computer vision problems is the heat diffusion process. Such Partial differential equations are continuous and have to be discretized by some techniques, mostly mathematical processes like finite differences or finite elements. The continuous domain is subdivide ..."
Abstract

Cited by 3 (3 self)
 Add to MetaCart
One physical process involved in many computer vision problems is the heat diffusion process. Such Partial differential equations are continuous and have to be discretized by some techniques, mostly mathematical processes like finite differences or finite elements. The continuous domain is subdivided into subdomains in which there is only one value. The diffusion equation comes from the energy conservation then it is valid on a whole domain. We use the global equation instead of discretize the PDE obtained by a limit process on this global equation. To encode these physical global values over pixels of different dimensions, we use a computational algebraic topology (CAT)based image model. This model has been proposed by Ziou and Allili and used for the deformation of curves and optical flow. It introduces the image support as a decomposition in terms of points, edges, surfaces, volumes, etc. Images of any dimensions can then be handled. After decomposing the physical principles of the heat transfer into basic laws, we recall the CATbased image model and use it to encode the basic laws. We then present experimental results for nonlinear graylevel diffusion for denoising, ensuring thin features preservation.
Tracking of Image Intensities Based on Optical Flow: An Evaluation of Nonlinear Diffusion Process
 In Second IEEE International Symposium on Signal Processing and Information Technology
, 2002
"... Image smoothing based on anisotropic diffusion provides better performance than classical linear filtering. One difficulty is to compare and evaluate the different models proposed. Since the basic model arises from fluid mechanics, we propose to track anisotropic diffusion using an optical flow tech ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Image smoothing based on anisotropic diffusion provides better performance than classical linear filtering. One difficulty is to compare and evaluate the different models proposed. Since the basic model arises from fluid mechanics, we propose to track anisotropic diffusion using an optical flow technique through a new physics based image model. Consequently to this, we propose a new choice for the conductance function which is easier to control. We demonstrate with the optical flow tracking that this function has a better behavior concerning the gradient threshold than the known conductance functions. 1
A Global Cat Approach For Graylevel Diffusion
"... This paper proposes an alternative to partial differential equations (PDEs) for the solution of diffusion (Perona and Malik scheme), using the heat transfer problem. Traditionally, the method for solving such physicsbased problems is to discretize and solve a PDE by a mathematical process. We propo ..."
Abstract
 Add to MetaCart
This paper proposes an alternative to partial differential equations (PDEs) for the solution of diffusion (Perona and Malik scheme), using the heat transfer problem. Traditionally, the method for solving such physicsbased problems is to discretize and solve a PDE by a mathematical process. We propose to use the global heat equation and decompose it into simpler laws. Some of these laws admit an exact global version since they arise from conservation principles while the assumptions on the others can be made wisely, taking into account knowledge about the problem. A computational algebraic topologybased image model allows us to write directly discrete equations. The numerical scheme is derived in a straightforward way from the problem modeled. It thus provides a physical explanation of each solving step in the solution. Finally, we present results for non linear diffusion.
Tracking of the Left Ventricle in Contrast Enhanced Echocardiography by Anisotropic Filtering and Active Contours Algorithm
"... Contrast agents together with second harmonic imaging improve endocardial border detection in echo images. Aim of the study was to assess whether non linear anisotropic filtering after segmentation of gradient vector flow (GVF) based active contour algorithm can be applied on contrast echo images fo ..."
Abstract
 Add to MetaCart
Contrast agents together with second harmonic imaging improve endocardial border detection in echo images. Aim of the study was to assess whether non linear anisotropic filtering after segmentation of gradient vector flow (GVF) based active contour algorithm can be applied on contrast echo images for automatic detection of left ventricular (LV) contours. LV volumes were measured in 20 subjects by cine Magnetic Resonance (MR) and transthoracic second harmonic echo after intravenous administration of the contrast agent Levovist. Contrast echo images were analyzed both manually and automatically; MR images were automatically analyzed. LV volumes were calculated according to Simpson’s rule. A good correlation with MR volumes was found for manual LV volumes (r.93 and.92 for end diastolic and end systolic volumes, respectively, p <.05). A still good correlation was found for the automatically assessed volumes in both phases (r=.90 and.91, p <05). Thus automatic endocardial border delineation by anisotropic filtering and GVF based active contour algorithm is feasible on contrast echo images. This method may represent an approach for automatic endocardial border analysis in ultrasonic imaging. 1.
Pattern Recognition] (]]]])]]]–]]] Contents lists available at SciVerse ScienceDirect Pattern Recognition
"... journal homepage: www.elsevier.com/locate/pr On the choice of the parameters for anisotropic diffusion in image processing ..."
Abstract
 Add to MetaCart
journal homepage: www.elsevier.com/locate/pr On the choice of the parameters for anisotropic diffusion in image processing