Robert Collins. Model Acquisition using Stochastic Projective Geometry. PhD thesis, University Massachusetts, September  Home/Search Document Details and Download
Summary Related Articles Check |
This paper is cited in the following contexts:
Scalable Extrinsic Calibration of Omni-Directional Image Networks - Antone, Teller (2002) (Correct)
....Methods that use local constraints alone (e.g. FZ98] are prone to bias and error accumulation. None of these methods have been demonstrated for noisy observations, significant occlusion and clutter, varying illumination, and wide baselines. Many researchers have used the Hough transform (e.g. [Bar83, Col93, LMLK94, Shu99]) for e#cient processing of large feature sets. However, HT accuracy is inherently limited by discretization, and no principled characterization of uncertainty has been demonstrated in this setting. The HT has been used to initialize continuous space VP estimation [Col93] This method uses ....
....[Bar83, Col93, LMLK94, Shu99] for e#cient processing of large feature sets. However, HT accuracy is inherently limited by discretization, and no principled characterization of uncertainty has been demonstrated in this setting. The HT has been used to initialize continuous space VP estimation [Col93]. This method uses deterministic clustering and outlier rejection, biasing the resulting VP estimates. 6.3 Unknown Correspondence Determining correspondence is a central problem in computer vision. A variety of interactive tools rely on a human operator to match features across images [BB95, ....
Robert T. Collins. Model Acquisition using Stochastic Projective Geometry. PhD thesis, University of Massachusetts, September 1993.
Robust Camera Pose Recovery Using Stochastic Geometry - Antone (2001) (2 citations) (Correct)
....transformation but which practical inference but impossible. Prentice [Pre84] also shown that maximum likelihood estimates spherical obtained without explicit distributions. Spherical uncertainty applied to projective features, stochastic projective geometry , treated extensively Collins [CW90, Col93], who demonstrates that Bingham s distribution, described in next section, a suitable choice for projective inference problems that overcomes many shortcomings other stochastic models. 3.3.2 Bingham s Distribution Exponential distributions have many appealing properties [Ber79] the least which ....
....measurement i to corrupted Bingham process with parameter matrix . Unlike traditional errors in variables approaches (e.g. MM00] underlying noise model neither additive nor Gaussian; however, since Bing ham s distribution exponential, analogous errors in variables methods formulated the sphere [Col93]. # represent random variable resulting from the fusion data Using Bayesian arguments from 3.2.2, posterior density be written p(# X) 1 c p(#) 3 31) where normalizing marginal probability ) which does depend prior density is assumed Bingham form, with parameter matrix . problem is now determine ....
[Article contains additional citation context not shown here]
Robert Collins. Model Acquisition using Stochastic Projective Geometry. PhD thesis, University Massachusetts, September
Scalable Rotational Registration of Omni-Directional Image.. - Antone, Teller (2001) (Correct)
....addresses this problem [30] However, the approach proposed in [30] expends quadratic time in the number of line features. Clustering nearby line intersections, for example by deterministic k means algorithms, works well when the number of outliers is small, but is not robust to noise. Collins [11] proposes an elegant use of the HT for reliable detection and clustering, followed by a more careful projective inference approach for accurate estimation of each VP. His overall clustering approach is deterministic, however, and uses a hard threshold to reject outliers, biasing the resulting ....
Robert T. Collins. Model Acquisition using Stochastic Projective Geometry. PhD thesis, UMASS, Sep. 1993.
Automatic Recovery of Camera . . . - Antone, al. (2000) (Correct)
....although features are detected in the Euclidean space of a planar image, they are treated as projective quantities determined solely by their 3 D ray directions. A Euclidean uncertainty model such as a Gaussian distribution cannot be applied to such quantities, so we utilize Bingham s distribution [5, 10], which exhibits antipodal symmetry and can describe a wide variety of shapes on the sphere (e.g. uniform, bipolar, and equatorial) Bingham s distribution is given by (1) where is a real symmetric parameter matrix, is a normalizing coefficient, and is a 3 D vector with . Many analogies can be ....
Collins, R. T. "Model Acquisition using Stochastic Projective Geometry". PhD thesis, University of Massachusetts, 1993.
Matthew E. Antone - Bachelor Science Electrical (Correct)
No context found.
Robert Collins. Model Acquisition using Stochastic Projective Geometry. PhD thesis, University Massachusetts, September
3-D Reconstruction of Static Human Body Shape from Image Sequence - Remondino (2004) (Correct)
No context found.
R.T. Collins, Model Acquisition using stochastic projective geometry, PhD Thesis, Computer Science Dep., University of Massachusetts, 1993.
3D Reconstruction of Static Human Body with a Digital Camera - Remondino (Correct)
No context found.
Collins, R.T., Model acquisition using stochastic projective geometry, PhD thesis, Computer Science Department, Massachussetts University, 1993
Human Body Reconstruction from Image Sequences - Remondino (2002) (Correct)
No context found.
Collins, R.T.: Model acquisition using stochastic projective geometry. PhD thesis, Computer Science Dep., University of Massachusetts, 1993
Exploiting Quaternions to Support Expressive Interactive.. - Johnson (2003) (Correct)
No context found.
Robert T. Collins. Model Acquisition Using Stochastic Projective Geometry. PhD thesis, University of Massachusetts, 1993.
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