An algorithm for automatically comparing images of two-dimensional electrophoretic gels is described. 2-D gels are useful in a wide range of biomedical problems. In particular, the 2-D gel comparison algorithm is bein...
详细信息
ISBN:
(纸本)0818606339
An algorithm for automatically comparing images of two-dimensional electrophoretic gels is described. 2-D gels are useful in a wide range of biomedical problems. In particular, the 2-D gel comparison algorithm is being used to empirically estimate the rate of mutation in human germinal cells. This application requires that the algorithm accurately describe both matches and mismatches between gel images. One issue that is addressed concerns the ways to distinguish between image differences that are real and those that result from noise. Another issue is that of maintaining consistency between match (or mismatch) hypotheses. These issues are also of interest to the more general image matching problem.
The optic flow problem, that is, the determination of 3-D structure from a sequence of 2-D images, is considered. A first stage in the solution has been the calculation of an image-plane velocity flow map from the ima...
详细信息
ISBN:
(纸本)0818606339
The optic flow problem, that is, the determination of 3-D structure from a sequence of 2-D images, is considered. A first stage in the solution has been the calculation of an image-plane velocity flow map from the image sequence. Correspondence between images is achieved through motion invariants, assumed to hold on the image plane. However, the image plane motion invariant is an approximation to a true '3-D world' motion invariant, and therefore leads to difficulties in situations where the scene structure and motion are truly three dimensional. It is shown how these difficulties are overcome by formulating the correspondence problem with true motion invariants. The result is an algorithm for the determination of scene structure directly from the image plane data without the flow map.
The moments of a picture are useful in deriving information about size and shape. Polynomial representations of pictures reduce storage requirements, suppress noise, and are easy to analyze. These two methods of repre...
详细信息
ISBN:
(纸本)0818606339
The moments of a picture are useful in deriving information about size and shape. Polynomial representations of pictures reduce storage requirements, suppress noise, and are easy to analyze. These two methods of representing pictures are related by a linear transformation that depends on the set over which the representations are computed. The transformation is derived here when the underlying set is a circle, a regular k-gon, and certain more general figures. Explicit formulas are derived for computing the quadratic polynomial coefficients from quadratic moments. A simple measure of the quality of a polynomial approximation is presented.
A new parameterization for lines in images is presented with application to the Hough transform. The modified Hough (Muff) transform has several implementational advantages over the conventional rho parameterization. ...
详细信息
ISBN:
(纸本)0818606339
A new parameterization for lines in images is presented with application to the Hough transform. The modified Hough (Muff) transform has several implementational advantages over the conventional rho parameterization. The Muff transform parameter space is better suited to computer graphics line drawing routines. The Muff transform requires no transcendental function calls or table lookup. The relation between the tessellation of the parameter space and the resolution of the lines represented is discussed. The shape of the Muff space is amenable to compaction into a rectangular array. The implementation of the Muff transform is presented.
The temporal covariance method (TCM) has been proposed as a motion detecting model with direction-sensitive units for human vision simulation. It consists of two subunits adjusted to opposite directions and a linear o...
详细信息
ISBN:
(纸本)0818606339
The temporal covariance method (TCM) has been proposed as a motion detecting model with direction-sensitive units for human vision simulation. It consists of two subunits adjusted to opposite directions and a linear operation that takes the difference between these direction-sensitive subunits. The authors describe how this model can be associated with a functional approximation of the spatiotemporal grayvalue variations to compute the local displacement vector field components.
A system for segmenting images based on edges is presented. The system, named CONTAM, is designed to segment images for which no assumptions can be made regarding such image properties irradiance, texture, or number a...
详细信息
ISBN:
(纸本)0818606339
A system for segmenting images based on edges is presented. The system, named CONTAM, is designed to segment images for which no assumptions can be made regarding such image properties irradiance, texture, or number and shape of objects. The system is applied to finding contamination in images of certain surface materials. First, a review of some well-known segmentation techniques is given in terms of their applicability to the problem at hand. Then, an overviews of CONTAM is given. The various modules are discussed, and new algorithms developed for the segmentation are described in detail. Finally, the system is justified by showing its performance on various kinds of real materials.
The authors evaluate the advantage of using an accurate initial guess in one of the ART (algebraic reconstruction techniques) algorithms, namely ART3. The model used is a binary image that represents a rectangle, tran...
详细信息
ISBN:
(纸本)0818606339
The authors evaluate the advantage of using an accurate initial guess in one of the ART (algebraic reconstruction techniques) algorithms, namely ART3. The model used is a binary image that represents a rectangle, translated or rotated. The initial guess is the same rectangle centered;this reflects the error due to misplacement of the object. The difference between the original and reconstructed rectangle is used as a measure of performance. The results show that when an accurate initial condition is used, there is a noticeable improvement in the first iterations. The difference, however decreases as the number of iterations increases. Thus a priori knowledge the reconstructed objects may be useful in decreasing the computation time.
The continuity equation is introduced as an improvement over the optical flow equation for determining velocity in images whose intensity is proportional to the density or to a projection of the density of some conser...
详细信息
ISBN:
(纸本)0818606339
The continuity equation is introduced as an improvement over the optical flow equation for determining velocity in images whose intensity is proportional to the density or to a projection of the density of some conserved quantity. Such 'density' images can be produced by both computerized axial tomography and X-ray imaging. A solution method for the continuity equation is presented that applies to any image and yields a velocity at every point in the image. The two models are compared, and each is used to determine the flow for a simple example. It is shown that the flows calculated by the two models can differ in both direction and magnitude and that the differences can be significant. Since the continuity equation, unlike the optical flow equation, is based on an actual physical flow, the continuity equation appears to be a more appropriate model for density images.
computer vision systems that attempt to extract cultural features from aerial imagery often are forced to interpret segmentations, where the actual features are broken into numerous segments or fragments. A method is ...
详细信息
ISBN:
(纸本)0818606339
computer vision systems that attempt to extract cultural features from aerial imagery often are forced to interpret segmentations, where the actual features are broken into numerous segments or fragments. A method is described for selecting sets of such fragments as candidates for alignment into a single region, as well as a procedure for generating new linear regions that are linked composites of the original sets of fragments. Portions of the composite region that lie between pairs of the original fragments are approximated with a spline. Therefore, the resulting composite region can be used to predict the areas in which to search for missing components of the cultural feature.
A closed curve in the plane can be described in several ways. It is shown that a representation in terms of radius of curvature versus tangent direction has certain advantages. In particular, a smoothed version of the...
详细信息
ISBN:
(纸本)0818606339
A closed curve in the plane can be described in several ways. It is shown that a representation in terms of radius of curvature versus tangent direction has certain advantages. In particular, a smoothed version of the closed curve can be computed easily using convolution. Similar filtering operations applied to some other representations of the curve do not guarantee that the result corresponds to a closed curve. In one case, where a closed curve is produced, it is smaller than the original. A description of a curve can be based on a sequence of smoothed versions of the curve. This is one reason why smoothing of closed curves is of interest.
暂无评论