Results 11  20
of
27
Camera pose revisited: New linear algorithms
 In 14„eme Congr„es Francophone de Reconnaissance des Formes et Intelligence Artificielle. Paper in French
, 2002
"... Abstract. Camera pose estimation is the problem of determining the position and orientation of an internally calibrated camera from known 3D reference points and their images. We briefly survey several existing methods for pose estimation, then introduce four new linear algorithms. The first three g ..."
Abstract

Cited by 12 (0 self)
 Add to MetaCart
Abstract. Camera pose estimation is the problem of determining the position and orientation of an internally calibrated camera from known 3D reference points and their images. We briefly survey several existing methods for pose estimation, then introduce four new linear algorithms. The first three give a unique linear solution from four points by SVD null space estimation. They are based on resultant matrices: the 24 × 24 method is the raw resultant matrix, and the 12 × 12 and 9 × 9 methods are compressed versions of this obtained by Gaussian elimination with pivoting on constant entries. The final method returns the four intrinsic solutions to the pose from 3 points problem. It is based on eigendecomposition of a 5 × 5 matrix. One advantage of all these methods is that they are simple to implement. In particular, the matrix entries are simple functions of the input data. Numerical experiments are given comparing the performance of the new algorithms with several existing algebraic and linear methods.
Generalised Measures of Reliability for Multiple Outliers
"... Abstract The application of the theory of reliability has become a fundamental part of measurement analysis, whether in order to optimise measurement systems so that they are resistant to the influence of outliers or in the postanalysis identification of outliers. However, the current theory of rel ..."
Abstract

Cited by 9 (7 self)
 Add to MetaCart
Abstract The application of the theory of reliability has become a fundamental part of measurement analysis, whether in order to optimise measurement systems so that they are resistant to the influence of outliers or in the postanalysis identification of outliers. However, the current theory of reliability is based on the assumption of a single outlier – an assumption that may not necessarily be the case. This paper extends reliability theory so that it can be applied to multiple outliers through the derivation of appropriate measures of reliability for multiple outliers. The measures of reliability covered include Minimal Detectable Biases, reliability numbers, controllability, and external reliability.
The Framework of Least Squares Template Matching
, 1998
"... The accurate registration of image series and exact matching of image parts is essential to numerous problems in computer vision. The framework presented in this paper is a generic matching algorithm suitable for many applications where feature extraction is difficult or inaccurate. Least squares te ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
The accurate registration of image series and exact matching of image parts is essential to numerous problems in computer vision. The framework presented in this paper is a generic matching algorithm suitable for many applications where feature extraction is difficult or inaccurate. Least squares template matching (LSM) is an areabased matching algorithm. It replaces the conventional multistage approach where feature detection is followed by thresholding, binarization and a discrete search. Thus, LSM does not depend on the extraction of binary (also called noniconic) image features. This is a very important advantage especially in lowcontrast and blurred imagery, where feature detection is mostly unreliable. Furthermore, unlike in most correlation methods, the optimum transformation is not searched by testing all possible cases, but approached using an optimization scheme. Assuming that a fair initial guess can be supplied, this is not only faster but also more accurate. Keywords ...
Optimal Fundamental Matrix Computation: Algorithm and Reliability Analysis
 Proc. 6th Symp. Sensing via Image Inf
, 2000
"... This paper presents an optimal linear algorithm for computing the fundamental matrix from corresponding points over two images under an assumed model, which admits independent Gaussian noise that is not necessarily isotropic or homogeneous. We derive a theoretical bound and demonstrate by experiment ..."
Abstract

Cited by 5 (2 self)
 Add to MetaCart
This paper presents an optimal linear algorithm for computing the fundamental matrix from corresponding points over two images under an assumed model, which admits independent Gaussian noise that is not necessarily isotropic or homogeneous. We derive a theoretical bound and demonstrate by experiments that our algorithm delivers results in the vicinity of the bound. Simulated and realimage examples are shown to observe the performance of our algorithm and illustrate the reliability evaluation process. 1.
Unified target detection and tracking using motion coherence
 in Motion, 2005
, 2005
"... This paper presents a unified approach to adaptive target detection and tracking. The unifying concept is “coherent motion energy”, a measure of the extent to which a single motion dominates local spatiotemporal structure. There are three major components to the approach. First, a multiresolution an ..."
Abstract

Cited by 4 (1 self)
 Add to MetaCart
This paper presents a unified approach to adaptive target detection and tracking. The unifying concept is “coherent motion energy”, a measure of the extent to which a single motion dominates local spatiotemporal structure. There are three major components to the approach. First, a multiresolution analysis of coherent motion energy is used to detect salient dynamic targets. Second, a robust affine transformation estimator is used to recover frametoframe target motion across regions of interest defined by coherent motion. Third, a method of template adaptation based on coherent motion weighted goodness of match is used to drive automatic template update. Empirical evaluation of the approach shows the contribution of the various components and documents strong performance of the integrated whole. 1.
Deformable Areabased Template Matching with Application to Low Contrast Imagery
 In Proc. Intl
, 1998
"... This paper reviews the mathematical formalism of the least squares template matching (LSM) and presents a framework for automatic quality control of the resulting match. LSM is an iterative and areabased fitting method which replaces the conventional multistep procedure of feature extraction follo ..."
Abstract

Cited by 3 (0 self)
 Add to MetaCart
This paper reviews the mathematical formalism of the least squares template matching (LSM) and presents a framework for automatic quality control of the resulting match. LSM is an iterative and areabased fitting method which replaces the conventional multistep procedure of feature extraction followed by discrete parameter search. The technique is especially suitable for attaining very high precision or for processing lowcontrast, noisy and blurred imagery as it takes into account the full information of the image signal. The automatic quality controla component often missing in commonly used image matching methodsis achieved by selfdiagnostic measures supervising the iterative procedure. The development is driven by applying image analysis in order to control patient position in radiotherapy treatment. The exact positioning of patients during radiotherapy is essential for high precision treatment. Accurate information about patient position is gained by the automated matching...
Model Selection in Statistical Inference and Geometric Fitting
 Proc. 3rd Workshop on InformationBased Induction Sciences, July 2000, Izu
, 2000
"... : Taking line fitting to points in two dimensions as a typical example, we point out the inherent difference between statistical inference and geometric fitting. We describe their duality in the sense that the asymptotic properties of statistical inference in the limit of an infinite number of ob ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
: Taking line fitting to points in two dimensions as a typical example, we point out the inherent difference between statistical inference and geometric fitting. We describe their duality in the sense that the asymptotic properties of statistical inference in the limit of an infinite number of observations hold for geometric fitting in the limit of infinitesimal perturbations. We contrast stochastic model selection with geometric model selection and describe the difference between Akaike's AIC and the geometric AIC in their derivations. 1.
Stabilizing Image Mosaicing by the Geometric AIC
 Proc. 2nd Workshop on InformationBased Induction Sciences
, 1999
"... : The computation for image mosaicing using homographies is numerically unstable and causes large image distortions if the matching points are small in number and concentrated in a small region in each image. This instability stems from the fact that actual transformations of images are usually in a ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
: The computation for image mosaicing using homographies is numerically unstable and causes large image distortions if the matching points are small in number and concentrated in a small region in each image. This instability stems from the fact that actual transformations of images are usually in a small subgroup of the group of homographies. In this paper, we show that such undesirable distortions can be removed by model selection using the geometric AIC without introducing any empirical thresholds. We also present an optimization scheme based on the LevenbergMarquardt method and an analytical procedure for computing an initial guess. We demonstrate the effectiveness of our method by real image examples. Key words: homography, image mosaicing, panoramic image, statistical optimization, model selection, geometric AIC 1. Introduction Image mosaicing is a technique for integrating multiple images into one continuous image, often referred to as a panoramic image [11, 14, 15, 16]. This...
A Statistical model for the Reliability of Motion Tracking using Meannormalised Correlation
 In Workshop on Performance Characteristics of Vision Algorithms
, 1996
"... Recent statistical results on the distribution of extrema in a twodimensional Gaussian random field are used to predict the mean failure rate of a block matching algorithm which uses the meannormalised correlation as the matching metric. The failure rate depends on the noise model, the area of the ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
Recent statistical results on the distribution of extrema in a twodimensional Gaussian random field are used to predict the mean failure rate of a block matching algorithm which uses the meannormalised correlation as the matching metric. The failure rate depends on the noise model, the area of the correlation surface to be searched and the properties of the block to be matched. Reliability of tracking a particular block is characterised by the maximum possible search area predicted by the model for a given confidence of a correct match. Simulation results based on synthetic noise fields and blocks from real images are used to confirm that the mean failure rate of the algorithm matches the prediction obtained from the model. The analysis is then applied to real image sequences and the actual failure rate is found to be well within an order of magnitude of that predicted. These results have significant implications for the selection of image features for motion tracking and the design ...
Error Analysis of Feature Based Disparity Estimation
"... Abstract. For realtime disparity estimation from stereo images the coordinates of feature points are evaluated. This paper analyses the influence of camera noise on the accuracy of feature point coordinates of a feature point detector similar to the Harris Detector, modified for disparity estimatio ..."
Abstract

Cited by 1 (1 self)
 Add to MetaCart
Abstract. For realtime disparity estimation from stereo images the coordinates of feature points are evaluated. This paper analyses the influence of camera noise on the accuracy of feature point coordinates of a feature point detector similar to the Harris Detector, modified for disparity estimation. As a result the error variance of the horizontal coordinate of each feature point and the variance of each corresponding disparity value is calculated as a function of the image noise and the local intensity distribution. Disparities with insufficient accuracy can be discarded in order to ensure a given accuracy. The results of the error analysis are confirmed by experimental results. 1