Salient Point and Scale Detection by Minimum Likelihood

Kim Steenstrup Pedersen; Marco Loog; Pieter van Dorst

Salient Point and Scale Detection by Minimum Likelihood

Kim Steenstrup Pedersen, Marco Loog, Pieter van Dorst

Abstract

We propose a novel approach for detection of salient image points
and estimation of their intrinsic scales based on the fractional
Brownian image model. Under this model images are realisations of a
Gaussian random process on the plane. We define salient points as
points that have a locally unique image structure. Such points are
usually sparsely distributed in images and carry important
information about the image content. Locality is defined in terms of
the measurement scale of the filters used to describe the image
structure. Here we use partial derivatives of the image function
defined using linear scale space theory. We propose to detect
salient points and their intrinsic scale by detecting points in
scale-space that locally minimise the likelihood under the model.

Original language	English
Title of host publication	Gaussian processes in practice
Number of pages	14
Publisher	Microtome Publishing
Publication date	2007
Pages	59-72
Publication status	Published - 2007
Event	Gaussian Processes in Practice Workshop - Bletchley Park, United Kingdom Duration: 12 Jun 2006 → 13 Jun 2006

Conference

Conference	Gaussian Processes in Practice Workshop
Country/Territory	United Kingdom
City	Bletchley Park
Period	12/06/2006 → 13/06/2006

Series	JMLR: Workshop and Conference Proceedings
Number	1

Access to Document

http://jmlr.csail.mit.edu/proceedings/papers/v1/

Cite this

@inproceedings{0a2f4240b55111dcbee902004c4f4f50,

title = "Salient Point and Scale Detection by Minimum Likelihood",

abstract = "We propose a novel approach for detection of salient image pointsand estimation of their intrinsic scales based on the fractionalBrownian image model. Under this model images are realisations of aGaussian random process on the plane. We define salient points aspoints that have a locally unique image structure. Such points areusually sparsely distributed in images and carry importantinformation about the image content. Locality is defined in terms ofthe measurement scale of the filters used to describe the imagestructure. Here we use partial derivatives of the image functiondefined using linear scale space theory. We propose to detectsalient points and their intrinsic scale by detecting points inscale-space that locally minimise the likelihood under the model.",

author = "Pedersen, {Kim Steenstrup} and Marco Loog and {van Dorst}, Pieter",

year = "2007",

language = "English",

series = "JMLR: Workshop and Conference Proceedings",

publisher = "Microtome Publishing",

number = "1",

pages = "59--72",

booktitle = "Gaussian processes in practice",

note = "Gaussian Processes in Practice Workshop ; Conference date: 12-06-2006 Through 13-06-2006",

}

TY - GEN

T1 - Salient Point and Scale Detection by Minimum Likelihood

AU - Pedersen, Kim Steenstrup

AU - Loog, Marco

AU - van Dorst, Pieter

PY - 2007

Y1 - 2007

N2 - We propose a novel approach for detection of salient image pointsand estimation of their intrinsic scales based on the fractionalBrownian image model. Under this model images are realisations of aGaussian random process on the plane. We define salient points aspoints that have a locally unique image structure. Such points areusually sparsely distributed in images and carry importantinformation about the image content. Locality is defined in terms ofthe measurement scale of the filters used to describe the imagestructure. Here we use partial derivatives of the image functiondefined using linear scale space theory. We propose to detectsalient points and their intrinsic scale by detecting points inscale-space that locally minimise the likelihood under the model.

AB - We propose a novel approach for detection of salient image pointsand estimation of their intrinsic scales based on the fractionalBrownian image model. Under this model images are realisations of aGaussian random process on the plane. We define salient points aspoints that have a locally unique image structure. Such points areusually sparsely distributed in images and carry importantinformation about the image content. Locality is defined in terms ofthe measurement scale of the filters used to describe the imagestructure. Here we use partial derivatives of the image functiondefined using linear scale space theory. We propose to detectsalient points and their intrinsic scale by detecting points inscale-space that locally minimise the likelihood under the model.

M3 - Article in proceedings

T3 - JMLR: Workshop and Conference Proceedings

SP - 59

EP - 72

BT - Gaussian processes in practice

PB - Microtome Publishing

T2 - Gaussian Processes in Practice Workshop

Y2 - 12 June 2006 through 13 June 2006

ER -

Salient Point and Scale Detection by Minimum Likelihood

Abstract

Conference

Access to Document

Fingerprint

Cite this