## What value covariance information in estimating vision parameters (2001)

### Cached

### Download Links

Venue: | In Eighth IEEE International Conference on Computer Vision (ICCV 2001 |

Citations: | 11 - 3 self |

### BibTeX

@INPROCEEDINGS{Brooks01whatvalue,

author = {Michael J. Brooks and Wojciech Chojnacki and Darren Gawley and Anton Van Den Hengel},

title = {What value covariance information in estimating vision parameters},

booktitle = {In Eighth IEEE International Conference on Computer Vision (ICCV 2001},

year = {2001},

pages = {302--308},

publisher = {IEEE Computer Society}

}

### Years of Citing Articles

### OpenURL

### Abstract

Many parameter estimation methods used in computer vision are able to utilise covariance information describing the uncertainty of data measurements. This paper considers the value of this information to the estimation process when applied to measured image point locations. Covariance matrices are first described and a procedure is then outlined whereby covariances may be associated with image features located via a measurement process. An empirical study is made of the conditions under which covariance information enables generation of improved parameter estimates. Also explored is the extent to which the noise should be anisotropic and inhomogeneous if improvements are to be obtained over covariance-free methods. Critical in this is the devising of synthetic experiments under which noise conditions can be precisely controlled. Given that covariance information is, in itself, subject to estimation error, tests are also undertaken to determine the impact of imprecise covariance information upon the quality of parameter estimates. Finally, an experiment is carried out to assess the value of covariances in estimating the fundamental matrix from real images. 1.

### Citations

322 | Determining the epipolar geometry and its uncertainty: A review
- Zhang
- 1998
(Show Context)
Citation Context ...ion Many problems in computer vision can be couched in terms of parameter estimation from image-based measurements. Such problems arise in stereo vision, with the estimation of the fundamental matrix =-=[8, 12, 15, 18]-=-, in conic fitting, with the estimation of an ellipse’s coefficients [3,17], and in many other areas. Because such problems are typically very sensitive to noise, there has recently been considerable ... |

271 | R.B.: Direct least square fitting of ellipses
- Fitzgibbon, Pilu, et al.
- 1999
(Show Context)
Citation Context ... from image-based measurements. Such problems arise in stereo vision, with the estimation of the fundamental matrix [8, 12, 15, 18], in conic fitting, with the estimation of an ellipse’s coefficients =-=[3,17]-=-, and in many other areas. Because such problems are typically very sensitive to noise, there has recently been considerable interest in assessing how parameter estimation might be improved if additio... |

242 |
Statistical Optimization for Geometric Computation: Theory and Practice
- Kanatani
- 1996
(Show Context)
Citation Context ...itive to noise, there has recently been considerable interest in assessing how parameter estimation might be improved if additional information is available characterising the uncertainty of the data =-=[1, 4, 10, 11, 16]-=-. This uncertainty information is usually expressed in terms of covariance matrices. This paper investigates conditions under which the use of covariance matrices enables parameter estimates of improv... |

234 | The fundamental matrix: theory, algorithms, and stability analysis
- Luong, Faugeras
- 1996
(Show Context)
Citation Context ...ion Many problems in computer vision can be couched in terms of parameter estimation from image-based measurements. Such problems arise in stereo vision, with the estimation of the fundamental matrix =-=[8, 12, 15, 18]-=-, in conic fitting, with the estimation of an ellipse’s coefficients [3,17], and in many other areas. Because such problems are typically very sensitive to noise, there has recently been considerable ... |

220 | The development and comparison of robust methods for estimating the fundamental matrix
- Torr, Murray
- 1997
(Show Context)
Citation Context ...ion Many problems in computer vision can be couched in terms of parameter estimation from image-based measurements. Such problems arise in stereo vision, with the estimation of the fundamental matrix =-=[8, 12, 15, 18]-=-, in conic fitting, with the estimation of an ellipse’s coefficients [3,17], and in many other areas. Because such problems are typically very sensitive to noise, there has recently been considerable ... |

198 | Parameter estimation techniques: A tutorial with application to conic fitting
- Zhang
- 1997
(Show Context)
Citation Context ... from image-based measurements. Such problems arise in stereo vision, with the estimation of the fundamental matrix [8, 12, 15, 18], in conic fitting, with the estimation of an ellipse’s coefficients =-=[3,17]-=-, and in many other areas. Because such problems are typically very sensitive to noise, there has recently been considerable interest in assessing how parameter estimation might be improved if additio... |

118 |
defense of the eight-point algorithm
- In
- 1997
(Show Context)
Citation Context |

74 | Heteroscedastic regression in computer vision: problems with bilinear constraint
- Leedan, Meer
- 2000
(Show Context)
Citation Context ...itive to noise, there has recently been considerable interest in assessing how parameter estimation might be improved if additional information is available characterising the uncertainty of the data =-=[1, 4, 10, 11, 16]-=-. This uncertainty information is usually expressed in terms of covariance matrices. This paper investigates conditions under which the use of covariance matrices enables parameter estimates of improv... |

65 | A unified factorization algorithm for points, line segments and planes with uncertainty models
- Morris, Kanade
- 1998
(Show Context)
Citation Context ...� � � � ensures that greater uncertainty is associated with lower overall intensity change. has proved useful in such applications as optical flow computation, texture segmentation, and factorization =-=[2, 13, 14]-=-. Figure 2 shows the covariance ellipses generated when is computed from various corner data. Each ellipse has � � Covariances of similar form have been used for centroid estimation [6]. 4. Experiment... |

59 | On the fitting of surfaces to data with covariances
- Chojnacki, Brooks, et al.
- 1303
(Show Context)
Citation Context ...itive to noise, there has recently been considerable interest in assessing how parameter estimation might be improved if additional information is available characterising the uncertainty of the data =-=[1, 4, 10, 11, 16]-=-. This uncertainty information is usually expressed in terms of covariance matrices. This paper investigates conditions under which the use of covariance matrices enables parameter estimates of improv... |

37 | Determination of ego-motion from matched points
- Harris
- 1987
(Show Context)
Citation Context ...th real images Our final test of the usefulness of covariance information was whether or not it can improve the accuracy of an estimated fundamental matrix relating two real images. The Harris method =-=[7]-=- is commonly used as a means of extracting corner features from images for the purpose of stereo matching. This approach employs an ‘interest operator’ evaluated for each pixel, using image intensity ... |

18 | Optimal Estimation of Matching Constraints
- Triggs
- 1998
(Show Context)
Citation Context |

8 |
Image movement detection with reliability indices
- Ohta
- 1991
(Show Context)
Citation Context ...� � � � ensures that greater uncertainty is associated with lower overall intensity change. has proved useful in such applications as optical flow computation, texture segmentation, and factorization =-=[2, 13, 14]-=-. Figure 2 shows the covariance ellipses generated when is computed from various corner data. Each ellipse has � � Covariances of similar form have been used for centroid estimation [6]. 4. Experiment... |

5 |
On weighting and choosing constraints for optimally reconstructing the geometry of image triplets
- Forstner
- 2000
(Show Context)
Citation Context |

3 | Scale Characteristics of Local Autocovariances for Texture Segmentation
- FABER, FÖRSTNER
- 1999
(Show Context)
Citation Context ...� � � � ensures that greater uncertainty is associated with lower overall intensity change. has proved useful in such applications as optical flow computation, texture segmentation, and factorization =-=[2, 13, 14]-=-. Figure 2 shows the covariance ellipses generated when is computed from various corner data. Each ellipse has � � Covariances of similar form have been used for centroid estimation [6]. 4. Experiment... |