Salient Point and Scale Detection by Minimum Likelihood

Kim S. Pedersen, Marco Loog, Pieter Dorst
Gaussian Processes in Practice, PMLR 1:59-72, 2007.

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.

Cite this Paper


BibTeX
@InProceedings{pmlr-v1-pedersen07a, title = {Salient Point and Scale Detection by Minimum Likelihood}, author = {Pedersen, Kim S. and Loog, Marco and Dorst, Pieter}, booktitle = {Gaussian Processes in Practice}, pages = {59--72}, year = {2007}, editor = {Lawrence, Neil D. and Schwaighofer, Anton and Quiñonero Candela, Joaquin}, volume = {1}, series = {Proceedings of Machine Learning Research}, address = {Bletchley Park, UK}, month = {12--13 Jun}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v1/pedersen07a/pedersen07a.pdf}, url = {https://proceedings.mlr.press/v1/pedersen07a.html}, 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.} }
Endnote
%0 Conference Paper %T Salient Point and Scale Detection by Minimum Likelihood %A Kim S. Pedersen %A Marco Loog %A Pieter Dorst %B Gaussian Processes in Practice %C Proceedings of Machine Learning Research %D 2007 %E Neil D. Lawrence %E Anton Schwaighofer %E Joaquin Quiñonero Candela %F pmlr-v1-pedersen07a %I PMLR %P 59--72 %U https://proceedings.mlr.press/v1/pedersen07a.html %V 1 %X 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.
RIS
TY - CPAPER TI - Salient Point and Scale Detection by Minimum Likelihood AU - Kim S. Pedersen AU - Marco Loog AU - Pieter Dorst BT - Gaussian Processes in Practice DA - 2007/03/11 ED - Neil D. Lawrence ED - Anton Schwaighofer ED - Joaquin Quiñonero Candela ID - pmlr-v1-pedersen07a PB - PMLR DP - Proceedings of Machine Learning Research VL - 1 SP - 59 EP - 72 L1 - http://proceedings.mlr.press/v1/pedersen07a/pedersen07a.pdf UR - https://proceedings.mlr.press/v1/pedersen07a.html AB - 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. ER -
APA
Pedersen, K.S., Loog, M. & Dorst, P.. (2007). Salient Point and Scale Detection by Minimum Likelihood. Gaussian Processes in Practice, in Proceedings of Machine Learning Research 1:59-72 Available from https://proceedings.mlr.press/v1/pedersen07a.html.

Related Material