Wearable visual sensors provide views of the environment which are rich in information about the wearer's location, interactions and intentions. In the wearable domain, hand gesture recognition is the natural repl...
详细信息
ISBN:
(纸本)0769526861
Wearable visual sensors provide views of the environment which are rich in information about the wearer's location, interactions and intentions. In the wearable domain, hand gesture recognition is the natural replacement for keyboard input. We describe a framework combining a coarse-to-fine method for shape detection and a 3D tracking method that can identify pointing gestures and estimate their direction. the low computational complexity of both methods allows a real-time implementation that is applied to estimate the user's focus of attention and to control fast redirections of gaze of a wearable active camera. Experiments have demonstrated a level of robustness of this system in long and noisy image sequences.
this paper proposes a new method for image denoising with edge preservation, based on image multiresolution decomposition by a redundant wavelet transform. In our approach, edges are implicitly located and preserved i...
详细信息
ISBN:
(纸本)0769513301
this paper proposes a new method for image denoising with edge preservation, based on image multiresolution decomposition by a redundant wavelet transform. In our approach, edges are implicitly located and preserved in the wavelet domain, while noise is filtered out. At each resolution, the coefficients associated to noise and coefficients associated to edges are modeled by Gaussians, and a shrinkage function is assembled. the shrinkage functions are combined in consecutive resolution, and geometric constraints are applied to preserve edges that are not isolated. Finally, the inverse wavelet transform is applied to the modified coefficients. this method is adaptive, and performs well for images contaminated by natural and artificial noise.
fMRI is a non-invasive technique to evaluate brain areas related to a requested task. the event-related fMRI is one kind of its design, which permits not only spatial resolution, but also temporal correlation. this in...
详细信息
ISBN:
(纸本)0769513301
fMRI is a non-invasive technique to evaluate brain areas related to a requested task. the event-related fMRI is one kind of its design, which permits not only spatial resolution, but also temporal correlation. this information may be more easily understood by the clinicians, if showed in a dynamic way. the aim of this paper is to describe a fMRI visualization system to expose the brain images, in the best view to stand out the areas of activity as long as they appear. It will empower the technique, not only to impress the clinicians, but also to enhance the importance of temporal connectivities. this system will be a module of the CAVERNA virtual reality.
this paper presents an interactive algorithm for image denoising and segmentation. A global competition criterion is used to impose an order of processing on all image pixels. the smoothing step employs an evolution e...
详细信息
ISBN:
(纸本)0769513301
this paper presents an interactive algorithm for image denoising and segmentation. A global competition criterion is used to impose an order of processing on all image pixels. the smoothing step employs an evolution equation controlled by the local curvature to denoise the image while preserving the features. the interactive segmentation step requires the user to select one definitive seed per region. Region growing is initiated around provisory seeds, which are automatically detected, labeled and eventually merged by the algorithm. A simple merging mechanism is used to handle the topological transformations required to remove the image over-segmentation. It is shown that accurate and fast segmentation results can be achieved for gray and color images using this simple method. Extension to 3D images is straightforward and easily handled.
Topologically adaptable snakes, or simply T-snakes, are a standard tool for automatically identifying multiple segments in an image. this work introduces a novel approach for controlling the topology of a T-snake. It ...
详细信息
ISBN:
(纸本)0769522270
Topologically adaptable snakes, or simply T-snakes, are a standard tool for automatically identifying multiple segments in an image. this work introduces a novel approach for controlling the topology of a T-snake. It focuses on the loops formed by the so-called projected curve which is obtained at every stage of the snake evolution. the idea is to make that curve the image of a piecewise linear mapping of an adequate class. then, withthe help of an additional structure-the Loop-Tree-it is possible to decide in O(1) time whether the region enclosed by each loop has already been explored by the snake. this makes it possible to construct an enhanced algorithm for evolving T-snakes whose performance is assessed by means of statistics and examples.
this paper presents a new approach for interactive editing of multiresolution meshes. Editing operations (e.g. scaling and rotation) are confined to mesh regions of interest, called sub-meshes. Also, intuitive deforma...
详细信息
ISBN:
(纸本)0769522270
this paper presents a new approach for interactive editing of multiresolution meshes. Editing operations (e.g. scaling and rotation) are confined to mesh regions of interest, called sub-meshes. Also, intuitive deformation operations (e.g. tapering and twisting) as required for animation systems can be used as well, allowing the user to have control on the local deformations of mesh objects. For that, each sub-mesh is defined interactively by picking up the cells of its frontier. Alternatively, a sub-mesh can be delimited by the shortest path of edges that approximates the intersection between a mesh and a given cutting plane. In order to speed up editing operations, the mesh is first simplified, the target sub-mesh is edited, and finally the whole mesh is refined back to its original resolution. this is particularly important for large meshes.
A new three-dimensional reconstruction technique is presented that uses an integer control parameter, denoted beta, to produce a family of models from a given set of planar cross-sections. Parameter beta supports mult...
详细信息
ISBN:
(纸本)076951846X
A new three-dimensional reconstruction technique is presented that uses an integer control parameter, denoted beta, to produce a family of models from a given set of planar cross-sections. Parameter beta supports multiple choices for solving the correspondence problem, i.e., the problem of deciding which regions from two consecutive cross-sections must be connected into a single component. thus, unlike current reconstruction methods, the beta-connection algorithm enables the consideration of multiple alternatives when establishing region correspondence. In addition to this flexibility, which is useful in creating models with complex topologies, the algorithm produces PL-manifolds and respects the re-sampling condition, thus providing an interesting reconstruction solution for many practical visualization and numerical simulation applications.
this paper presents a novel approach to the real-time rendering of complex scenes problem. Up to the present date, a huge number of acceleration techniques have been proposed, although most are geared towards a specif...
详细信息
ISBN:
(纸本)0769513301
this paper presents a novel approach to the real-time rendering of complex scenes problem. Up to the present date, a huge number of acceleration techniques have been proposed, although most are geared towards a specific kind of scene. Instead of using a single, or a fixed set of rendering acceleration techniques, we propose the use of several, and to select the best one based on the current viewpoint. thus, dynamically adapting the rendering process to the contents of the scene, it is possible to take advantage of all these techniques when they are better suited, rendering scenes that would otherwise be too complex to display at interactive frame rates. We describe a framework capable of achieving this purpose, consisting on a pre-processor and an interactive rendering engine. the framework is geared towards interactive applications were a complex and large scene has to be rendered at interactive frame rates. Finally, results taken from our test implementation are given.
A good shape descriptor is necessary for automatically identifying landmarks on boundaries. Our method of boundary shape description is based on the notion of c-scale, which is a new local scale concept, defined at ea...
ISBN:
(纸本)9780769529967
A good shape descriptor is necessary for automatically identifying landmarks on boundaries. Our method of boundary shape description is based on the notion of c-scale, which is a new local scale concept, defined at each boundary element. From this representation we can extract special points of interest such as convex and concave corners, straight lines, circular segments, and inflection points. the results show that this method gives a complete description of shape and allows the automatic positioning of mathematical landmarks, which agree with our intuitive ideas of where landmarks may be defined. this method is applicable to spaces of any dimensionality, although we have focused in this paper on 2D shapes.
In most of video shot boundary detection algorithms, proposed in the literature, several parameters and thresholds have to be set in order to achieve good results. In this paper, to get rid of parameters and threshold...
详细信息
ISBN:
(纸本)0769526861
In most of video shot boundary detection algorithms, proposed in the literature, several parameters and thresholds have to be set in order to achieve good results. In this paper, to get rid of parameters and thresholds, we explore a supervised classification method for video shot segmentation. We transform the temporal segmentation into a class categorization issue. Our approach defines a uniform framework for combining different kinds of features extracted from the video. Our method does not require any pre-processing step to compensate motion or post-processing filtering to eliminate false detected transitions. the experiments, following strictly the TRECVID 2002 competition protocol, provide very good results dealing with a large amount of features thanks to our kernel-based SVM classification method.
暂无评论