D. J. Burr, "A dynamic model for image registration," Comput. Graphics Image Process., vol. 15, pp. 102--112, 1981.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....the search time can be significantly reduced without compromising recognition performance. 1997 Academic Press Deformable models have been used widely to characterize the shape variability of objects. Their main use has been to obtain a best fit match between the model and given data, as in [25, 10, 8, 9, 1, 11, 28]. This framework can also be extended to tracking objects over several frames (e.g. 5] and to object recognition [13, 23, 20] Deformable models are closely related to the work on active contour models or snakes in [18] A major problem with this framework is that the procedure for fitting a ....

D.J. Burr, A dynamic model for image registration. Cornput. Graphics Image Process. 15, 1981, 102-112.

....opposition to abrupt distortions which avoid unnatural deformities of the test image, even if they lead to a higher individual point to point similarity, on the other. One of the first elastic registration methods was the rubber mask technique [11] A more advanced approach was introduced in [3], where the basic elements of modern methods are already present. More recently, in [10] a number of refinements are proposed. The three phases which compose an elastic local registration method are: # Selection of the landmark or control points which will be used to command the deformation ....

D.J. Burr, "A Dynamic Model for Image Registration" Computer Graphics and Image Processing Vol. 15 pp. 102-112, 1981

....is correspondence between like features rather than exact matches. Widrow [21] also suggests the idea of using rubber templates to achieve fuzzy matches to a variety of natural objects and waveforms. Burr presents an iterative framework for computing elastic matches in dot and grey scale images [22] and line drawings [23] Using a coarse to fine matching strategy he shows how an image can be progressively deformed under the influence of misalignment force fields to fit another image. In a later version [24] global size and rotation adjustments were included. The method has been adapted to ....

....left its centre bar far from any inked pixels. We call this the beads in white space problem 7 . Motivated by research on snakes [31] another energy term, Ew , was defined to penalize beads in white space. This term is similar to the support measure [32] or the symmetric matching used in [22] and [23] 7 There is a small implicit penalty in that beads far from inked pixels are not available for accounting for inked pixels. A more elaborate generative model for both inked and non inked pixels is suggested in section V. 14 Ew = Gamma B X b=1 log N i X k=1 P kb (12) A bead ....

D. J. Burr, "A dynamic model for image registration", Comput. Graphics Image Process., vol. 15, pp. 102--112, 1981.

....might be unknown) the latter prove inapplicable in many cases: changes in luminosity cause the local intensities to vary, perspective deformations aoeect the gradient orientation. Most existing methods for matching images with weak constraints iteratively distort one image to t the second [2] [3] [7] The main drawbacks of these approaches are the need of a global preregistration process and their iterative implementation which results in large computational requirements. The approach we propose is aimed at exploiting to its best the geometric information provided by the contours ....

D.J. Burr. A dynamic model for image registration. Computer Graphics and Image, 15:102112, 1981.

....receptive field, that is, the kernel cannot be expressed as K(x, t) F(t)G(x) where F (t)and G(x) are functions which depend only on time and only on space, respectively. There are many models of direction selectivity(Barlow and Levick, 1965; Adelson and Bergen, 1985; Watson and Ahumada, 1985; Burr, 1981). In all of the models, the response of the cell is determined by receptive fields that have di#erent temporal response properties at di#erent spatial locations (i.e. spatiotemporal inseparable) This can be realized by the appropriate spatial positioning of the receptive fields, and the ....

Burr, D. (1981). A dynamic model for image registration. Computer Graphics and Image Processing, 15:102--112.

....more invariant to shape and scale, such as edges joined in a Y or a T , are used. In [Duda 73] it is suggested that a triangle be matched by first finding three separate lines and then determining if a triangle is indeed present. A better solution is offered by [Widrow 73] elaborated upon by [Burr 81] who introduces the rubber template, a template which can be locally distorted, so that information between local matches can be utilized. This is described in more detail in section 3.4 If the image is noisy, the peak of the correlation may not be clearly discernible. If the noise can be ....

....and electrocardiogram waveforms. The flexible template technique was implemented by defining specific parameters for the possible deformations in each problem domain. These were used to iteratively modify the tem34 plate until the best match was found. However, it was not until more recently [Burr 81] that automatic elastic registration methods were developed. Burr accomplished this by an iterative technique which depends on the local neighborhood whose size is progressively smaller with each iteration. At each iteration, the distance to the nearest neighbor in the complementary image is ....

D. J. Burr, "A Dynamic Model for Image Registration," Computer Graphics and Image Processing 15, 1981, pp102-112.

....are stored in a hash table which is indexed by curvature and torsion values. Curve points in the other image are compared with entries in the hash table and votes are accumulated for specific rigid body transformations. Moshfeghi [144] extended an elastic matching algorithm developed by Burr [23] to use contour information to register images of deformable anatomy. The algorithm uses an iterative Gaussian smoothed deformation model. A displacement vector is calculated between each point in one contour and its nearest line segment from the other contour. One image is deformed by computing a ....

D. J. Burr. A dynamic model for image registration. Comput. Vision Graphics Image Processing, 15:102--112, 1981.

....longer the loop of the 2 model would unfold to produce a final fit very much like that shown in (b) Deformable models have been used widely to characterize the shape variability of objects. Their main use has been to obtain a best fit match between the model and given data, as in [25] 10] [8], 9] 1] 11] and [28] This framework can also be extended to tracking objects over several frames (e.g. 5] and to object recognition ( 13] 23] 20] Deformable models are closely related to the work on active contour models or snakes in [18] A major problem with this framework is ....

D. J. Burr. A dynamic model for image registration. Comput. Graphics Image Process., 15:102--112, 1981.

....is correspondence between like features rather than exact matches. Widrow [20] also suggests the idea of using rubber templates to achieve fuzzy matches to a variety of natural objects and waveforms. Burr presents an iterative framework for computing elastic matches in dot and grey scale images [21] and line drawings [22] Using a coarse to fine matching strategy he shows how an image can be progressively deformed under the influence of misalignment force fields to fit another image. In a later version [23] global size and rotation adjustments were included. The method has been adapted to ....

....of the model to the image. Motivated by research on snakes [30] a simple approach to the beads in white space problem (section V.B) is to define another energy term, Ew to penalize beads spanning white space. This term is similar to the support measure [31] or the symmetric matching used in [21] and [22] Ew = Gamma B X b=1 log N I X k=1 P kb (21) A bead only makes a large contribution to this cost when all inked pixels are far from the bead. This energy term could be easily incorporated into the fitting procedure, but in the present system we simply use it as an additional term ....

D. J. Burr, "A dynamic model for image registration", Comput. Graphics Image Process., vol. 15, pp. 102--112, 1981.

No context found.

D. J. Burr, "A dynamic model for image registration," Comput. Graphics Image Process., vol. 15, pp. 102--112, 1981.

No context found.

Burr, D. J. (1981a). A dynamic model for image registration. Comput. Graphics Image Process., 15:102--112.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC