J. K. Laading, C. C. McCulloch, and V. E. Johnson. A hierarchical object deformation model applied to the digital chest radiograph. In The American Statistical Association Proceedings of the Section on Bayesian Statistical Science, Anaheim, California, 1997. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
A Hierarchical Feature Based Deformation Model Applied to.. - Jacob Laading Colin (1998) Self-citation (Laading Mcculloch Johnson) (Correct)
.... A l enforces some smoothness on the deformation, i.e. the marginal covariance between any pair of facets on a level l is a smooth decreasing function of their distance ( lj Gamma lj 0 ) in the reference image [14] This is unlike previous models in which only one non zero term per row was used[11]. Computational tractability is ensured by the choice of a normal hierarchical model. Feature distribution Let f be the vector of image derived feature values associated with the set of facets. Similarly, let OE be the corresponding vector of reference image feature values. Then given an image ....
....j ; OE j g from the image data can be specified in a number of ways; we have used either a quantile rescaled image intensity or a low scale image Laplacian. For choices of the functions g, scaled squared differences between f j and OE j (yielding independent normal distributions) have been used[15][11], and a local intensity regression, as described in the next subsection, has also been employed. 6 Laading et al. : Hierarchical Deformation Modeling for 4D SPECT 2.2 Application specifics The baseline shape model (3) captures the shape changes in the general case when little prior knowledge is ....
J. K. Laading, C. C. McCulloch, and V. E. Johnson. A hierarchical object deformation model applied to the digital chest radiograph. In The American Statistical Association Proceedings of the Section on Bayesian Statistical Science, Anaheim, California, 1997.
High-Level Image Understanding Via Bayesian Hierarchical Models - McCulloch (1998) Self-citation (Mcculloch) (Correct)
....for an observer to identify. This is discussed in section 2.6. Note that the layout of facets in the atlas image is completely arbitrary. Applications in this paper indicate that a grid should be used, however other schemes have been investigated. See, for instance, McCulloch et al. 1996) and Laading et al. 1997). 2.4 Data Model in this Paper In this section, a definition is given of the data model p(x j ; f j j; OE; used throughout this paper. Under independent sampling of facet properties (x j ; f j ) the model on J samples is Q J j=p p(x j ; f j j; OE; As mentioned in section 2.2, is ....
....that are expected to deform together within the class. This form of prior information is easily modeled by increasing the parent child weights between constituent facets in each object, and decreasing those connection across the object s boundary. One extreme approach has been investigated by Laading et al. 1997) in which certain parent child connections were set to zero, transforming the facet graph into a tree structure. More complex facet correlation structures could be modeled by treating all the parent child weights as unknown parameters and simulating from their (non standard) joint posterior ....
Laading, J. K., McCulloch, C. C. and Johnson, V. E. (1997) A hierarchical object deformation model applied to the digital chest radiograph. In The American Statistical Association Proceedings of the Section on Bayesian Statistical Science, Anaheim, California.
Image Feature Identification via Bayesian Hierarchical.. - McCulloch, Laading, Johnson (1997) Self-citation (Laading Mcculloch Johnson) (Correct)
....Future research will also address several generalizations of the model. First, template structure of the facet tree need not be generated by the method of Section 2. New trees will be constructed using the template image itself to explore different shape constraints on predicted facet locations [4]. Also, the structure of pS (x) will be generalized to have a conditional independence structure representable by a graph rather than a tree. By having more than one parent for each facet, spatially extended correlation structures between facets can be explored while keeping a high level of ....
J. K. Laading, C. McCulloch, and V. Johnson. A hierarchical object deformation model applied to the digital chest radiograph. In The American Statistical Association Proceedings of the Section on Bayesian Statistical Science, Anaheim, California, 1997.
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