Results 1  10
of
10
A PERFORMANCE EVALUATION OF LOCAL DESCRIPTORS
, 2005
"... In this paper we compare the performance of descriptors computed for local interest regions, as for example extracted by the HarrisAffine detector [32]. Many different descriptors have been proposed in the literature. However, it is unclear which descriptors are more appropriate and how their perfo ..."
Abstract

Cited by 1157 (38 self)
 Add to MetaCart
In this paper we compare the performance of descriptors computed for local interest regions, as for example extracted by the HarrisAffine detector [32]. Many different descriptors have been proposed in the literature. However, it is unclear which descriptors are more appropriate and how their performance depends on the interest region detector. The descriptors should be distinctive and at the same time robust to changes in viewing conditions as well as to errors of the detector. Our evaluation uses as criterion recall with respect to precision and is carried out for different image transformations. We compare shape context [3], steerable filters [12], PCASIFT [19], differential invariants [20], spin images [21], SIFT [26], complex filters [37], moment invariants [43], and crosscorrelation for different types of interest regions. We also propose an extension of the SIFT descriptor, and show that it outperforms the original method. Furthermore, we observe that the ranking of the descriptors is mostly independent of the interest region detector and that the SIFT based descriptors perform best. Moments and steerable filters show the best performance among the low dimensional descriptors.
General Intensity Transformations and Differential Invariants
, 1994
"... We consider the group of invertible image grayvalue transformations and propose a generating equation for a complete set of differential grayvalue invariants up to any order. Such invariants describe the image’s geometrical structure independent of how its grayvalues are mapped (contrast or brig ..."
Abstract

Cited by 28 (3 self)
 Add to MetaCart
We consider the group of invertible image grayvalue transformations and propose a generating equation for a complete set of differential grayvalue invariants up to any order. Such invariants describe the image’s geometrical structure independent of how its grayvalues are mapped (contrast or brightness adjustments).
Pattern Recognition in Images By Symmetries and Coordinate Transformations
, 1997
"... A theory for detecting general curve families by means of symmetry measurements in the coordinate transformed originals is presented. Symmetries are modeled by isogray curves of conjugate harmonic function pairs which also define the coordinate transformations. Harmonic function pair coordinates re ..."
Abstract

Cited by 23 (4 self)
 Add to MetaCart
A theory for detecting general curve families by means of symmetry measurements in the coordinate transformed originals is presented. Symmetries are modeled by isogray curves of conjugate harmonic function pairs which also define the coordinate transformations. Harmonic function pair coordinates render the target curve patterns as parallel lines, which is defined here as linear symmetry. Detecting these lines, or generalized linear symmetry fitting as it will be called, corresponds to finding invariants of Lie groups of transformations. A technique based on least square error minimization for estimating the invariance parameters is presented. It uses the Lie infinitesimal operators to construct feature extraction methods that are efficient and simple to implement. The technique, which is shown to be an extension of the generalized Hough transform, enables detection by voting and accumulating evidence for the searched pattern. In this approach complex valued votes are permitted, where ...
Matching by local invariants
, 1995
"... apport de recherche ISSN 02496399Matching by local invariants ..."
Abstract

Cited by 14 (2 self)
 Add to MetaCart
apport de recherche ISSN 02496399Matching by local invariants
Symmetry Derivatives of Gaussians Illustrated by Cross Tracking
, 2001
"... We propose a family of complex differential operators, symmetry derivatives, for pattern recognition in images. We present three theorems on their properties as applied to Gaussians. These show that all orders of symmetry derivatives of Gaussians yield compact expressions obtained by replacing the o ..."
Abstract

Cited by 4 (4 self)
 Add to MetaCart
We propose a family of complex differential operators, symmetry derivatives, for pattern recognition in images. We present three theorems on their properties as applied to Gaussians. These show that all orders of symmetry derivatives of Gaussians yield compact expressions obtained by replacing the original differential polynomial with an ordinary polynomial. Just like Gaussians, the symmetry derivatives of Gaussians are (form) invariant to Fourier transform, that is they are rescaled versions of the original. As a result, the symmetry derivatives of Gaussians are closed under the convolution operator, i.e. they map on a member of the family when convolved with each other. Since Gaussians are utilized extensively in image processing, the revealed properties have practical consequences, e.g. when designing filters and filtering schemes that are unbiased w.r.t. orientation (isotropic). A use of these results is illustrated by an application: tracking the cross markers in long image sequences from vehicle crash tests. The implementation and the results of this application are discussed in terms of the theorems presented, along with conclusions. 1
Isophotes: the Key to Tractable Local Shading Analysis
"... . When studying the formation of intensity image of an opaque surface, a natural question arises: how does the local surface structure project on the local image structure? We consider this question for the case of a single and monocular intensity image. In this paper, we discuss the the general pro ..."
Abstract

Cited by 2 (0 self)
 Add to MetaCart
. When studying the formation of intensity image of an opaque surface, a natural question arises: how does the local surface structure project on the local image structure? We consider this question for the case of a single and monocular intensity image. In this paper, we discuss the the general problem of local shading analysis (LSA) of an arbitrary C 2 surface. We reformulate the problem so that it is well tractable without any restrictive assumptions about the surface. We also contribute to the discussion about the LSA ambiguity problem by identifying image structures where the local surface interpretation may only change. 1 Introduction Local Shading Analysis (LSA) aims at looking for a direct relation between differential surface structure and the local structure of its intensity image. By surface it is meant surface in the small, i.e., defined in some open neighbourhood of one of its points. The structure is described by an mth order local model. The general advantage of LSA...
Harris Feature Vector Descriptor (HFVD)
"... A new image feature called Harris feature vector is defined in this paper, which effectively describes the image gradient distribution. By computing the mean and the standard deviation of the Harris feature vector in key point neighborhood, a novel descriptor for key points matching is constructed, ..."
Abstract

Cited by 1 (0 self)
 Add to MetaCart
A new image feature called Harris feature vector is defined in this paper, which effectively describes the image gradient distribution. By computing the mean and the standard deviation of the Harris feature vector in key point neighborhood, a novel descriptor for key points matching is constructed, which is invariant to image rigid transformation and linear intensity change. Experimental evidence suggests that the novel descriptor has a good adaptability to slight view point changing, JPEG compression as well as nonlinear change of intensity. 1.
A Performance Evaluation of Local Descriptors
, 2003
"... In this paper we compare the performance of interest point descriptors. Many different descriptors have been proposed in the literature. However, it is unclear which descriptors are more appropriate and how their performance depends on the interest point detector. The descriptors should be distincti ..."
Abstract
 Add to MetaCart
In this paper we compare the performance of interest point descriptors. Many different descriptors have been proposed in the literature. However, it is unclear which descriptors are more appropriate and how their performance depends on the interest point detector. The descriptors should be distinctive and at the same time robust to changes in viewing conditions as well as to errors of the point detector. Our evaluation uses as criterion detection rate with respect to false positive rate and is carried out for different image transformations. We compare SIFT descriptors [11], steerable filters [5], differential invariants [10], complex filters [17], moment invariants [21] and crosscorrelation for different types of interest points [8, 11, 13, 14]. In this evaluation, we observe that the ranking of the descriptors does not depend on the point detector and that SIFT descriptors perform best. Steerable filters come second ; they can be considered a good choice given the low dimensionality.
Abstract Distinctive Feature Analysis of Natural Landmarks as a Front end for SLAM applications
"... This paper presents a method for extracting distinctive textural features from images taken from natural scenes. The aim is to use natural landmarks for navigation in an unexplored environment. Natural features are all different and complex in shape. To be able to use them for navigation, informativ ..."
Abstract
 Add to MetaCart
This paper presents a method for extracting distinctive textural features from images taken from natural scenes. The aim is to use natural landmarks for navigation in an unexplored environment. Natural features are all different and complex in shape. To be able to use them for navigation, informative representation of these features and a careful selection process is required. The present method is termed as ‘Distinctive Texture Analysis’. It has three parts. Firstly, a method of selecting Interest Points from the filtered images is presented. Secondly, Texture Analysis of the local properties of Interest Points are applied and stored as descriptors. Thirdly, to reduce the number of landmarks selected for storage and comparison purposes, Distinctness Analysis is applied. Current results have shown that the most distinctive features as concurred by viewing the simple images are able to be selected and correctly matched. Results provided for the complex underwater images illustrated the difficulty and limitation. However, when this method is applied with multiple numbers of landmarks such that correlation of landmark positions is considered, certainty for SLAM can increase. Future works can include consideration of such correlation.