A parallel processing.structure and its control scheme are proposed for fast binary image normalization. The structure is a 4-neighbor-connected mesh of simple processing.elements (PEs). Each PE has a queue dedicated ...
详细信息
ISBN:
(纸本)0818606339
A parallel processing.structure and its control scheme are proposed for fast binary image normalization. The structure is a 4-neighbor-connected mesh of simple processing.elements (PEs). Each PE has a queue dedicated to the temporary storage of data. The normalization process is decomposed into three procedures: translation, scaling, and rotation. In each procedure the mapping of a pixel from its original position to its destination is controlled by the repetitive application of basic flow controls. This approach considerably simplifies the control and data routing algorithms. Normalization is completed with parallel boundary reconnection and filling algorithms. The results of a computer simulation are also provided.
A pixel-based maximum-likelihood texture segmentation algorithm is presented. Assuming that the textures in the given image are generated by Gaussian Markov random field models of known structure and parameters, the a...
详细信息
ISBN:
(纸本)0818606339
A pixel-based maximum-likelihood texture segmentation algorithm is presented. Assuming that the textures in the given image are generated by Gaussian Markov random field models of known structure and parameters, the authors maximize the likelihood that the particular image data have occurred. The resulting algorithm involves estimating the sample correlations over a window centered at the pixel being labeled and can be implemented using parallel computations.
Recent research in imageprocessing.has produced data structures for representing pictorial information that are not only space-efficient, but are also translation-invariant. Recently, a new data structure called TID ...
详细信息
ISBN:
(纸本)0818606339
Recent research in imageprocessing.has produced data structures for representing pictorial information that are not only space-efficient, but are also translation-invariant. Recently, a new data structure called TID (translation invariant data structure) was introduced. TID is not sensitive to the placement of the region and is translation-invariant. The authors compare the space-efficiency and sensitivity of TID to other data structures.
A method is described for using a sequence of views of a scene to construct a representation of the world in terms of space occupied by objects and space that is empty. The world is represented by an octree, and it is...
详细信息
ISBN:
(纸本)0818606339
A method is described for using a sequence of views of a scene to construct a representation of the world in terms of space occupied by objects and space that is empty. The world is represented by an octree, and it is assumed that each view is taken from an arbitrary, but known, location. The spatial representation is part of the description of the world required for sensory-interactive control of a robot manipulator. It is useful for trajectory planning and provides a spatial index of the objects in the world.
A description is given of an approach to the identification of shaded segments in an image providing a model for the curved surface that generated the segment. An algorithm that obtains the segments is carefully expla...
详细信息
ISBN:
(纸本)0818606339
A description is given of an approach to the identification of shaded segments in an image providing a model for the curved surface that generated the segment. An algorithm that obtains the segments is carefully explained. The output of the algorithm is a list of the image segments where shading was identified along with the geometric and curved surface model parameters of each segment. The image gradient (BxBy) map, a two-dimensional plot of the distribution of image points in the image gradient space, is utilized as a tool in the development of the algorithm. The algorithm provides a description of images in terms of components which correspond to curved surfaces in scenes. Initial results have demonstrated the feasibility and reliability of the algorithm.
One approach to segmenting an image into nearly homogeneous regions, separated by smooth boundaries, consists in finding the minimum of a functional that is formed by combining the length of the hypothetical boundary,...
详细信息
ISBN:
(纸本)0818606339
One approach to segmenting an image into nearly homogeneous regions, separated by smooth boundaries, consists in finding the minimum of a functional that is formed by combining the length of the hypothetical boundary, the gradient of a smoothed version of the image, and the difference of the true and smoothed images. The authors propose a method of finding a pseudominimum by solving the Euler-Lagrange equations for the boundaries, under the assumption that the boundaries are not too curved or too close, and then using hill-climbing. The method is worked out here in the simpler case of segmenting an interval into subintervals on which a 1-dimensional signal is roughly constant.
作者：
Chung, K.L.Natl Taiwan Univ
Dep of Information Engineering Taiwan Natl Taiwan Univ Dep of Information Engineering Taiwan
The author describes an efficient asynchronous scheme, called FCLC (first checking, last computing) pyramid architecture, for bottom-up image analysis. The scheme which is intended to compensate for the clock skews pr...
详细信息
ISBN:
(纸本)0818606339
The author describes an efficient asynchronous scheme, called FCLC (first checking, last computing) pyramid architecture, for bottom-up image analysis. The scheme which is intended to compensate for the clock skews problem existing in many synchronous systems, not only allows high performance to be achieved in pipelining fashion but also utilizes image compression, especially for single image tasks. Thus the computation time of the image algorithm is image dependent. The operation of the FCLC pyramid architecture is illustrated by examples of standard algorithms for interior-based computations (e. g. , area) and border-based computations of local properties (e. g. , perimeter). Finally FCLC is compared with other schemes.
作者：
Roberts, K.S.Columbia Univ
Dep of Computer Science New York NY USA Columbia Univ Dep of Computer Science New York NY USA
The equivalence problem for shape descriptions is that a single three-dimensional shape may have several different descriptions. The slant theorem (S. A. Schafer, 1983) for equivalent generalized cylinder descriptions...
详细信息
ISBN:
(纸本)0818606339
The equivalence problem for shape descriptions is that a single three-dimensional shape may have several different descriptions. The slant theorem (S. A. Schafer, 1983) for equivalent generalized cylinder descriptions was proved under the restrictions that the same radius function and the same axis be used for all the descriptions. A proof is given that the theorem still holds when the same radius function condition is removed. It does not hold when the same axis condition is removed. The ellipsoid is a counterexample.
The problem of defining a goodness measure for object labeling problems is addressed. The optimal probabilistic measure is derived but shown to be impractical for realistic problems. An alternative consistency measure...
详细信息
ISBN:
(纸本)0818606339
The problem of defining a goodness measure for object labeling problems is addressed. The optimal probabilistic measure is derived but shown to be impractical for realistic problems. An alternative consistency measure is suggested, which is based on the accuracy with which each object can estimate its label. This measure is shown to be a generalization of the crisp idea of consistency and can be computed using statistical information about the problem. Results are shown on a triangle labeling example and on gray-level picture segmentation and border detection.
One-dimensional local features are the basic primitives in most image analysis. These primitives can be detected (measured) with operators that deliver both magnitude and orientation response. It is shown how to desig...
详细信息
ISBN:
(纸本)0818606339
One-dimensional local features are the basic primitives in most image analysis. These primitives can be detected (measured) with operators that deliver both magnitude and orientation response. It is shown how to design such operators matched to a given profile of a specific 1D-feature. The procedure consists of one forward one-dimensional Fourier transform followed by one inverse Hankel transform. The result suffices to design a fully rotation-invariant convolution kernel matched to an edge, a ramp, or a line function of arbitrary profile shape.
暂无评论