Active shape model: Difference between revisions

Revision as of 20:07, 23 May 2006

Latest revision as of 10:28, 5 October 2023

(47 intermediate revisions by 31 users not shown)

Line 1:

{{Short description|Statistical model of an object's shape}}'''Active shape models''' ('''ASMs''') are [[statistical model]]s of the [[shape]] of objects which iteratively deform to fit to an example of the object in a new image, developed by Tim Cootes and Chris Taylor in 1995.<ref name=Cootes>{{cite journal| author=T.F. Cootes and C.J. Taylor and D.H. Cooper and J. Graham| title=Active shape models - their training and application| journal=Computer Vision and Image Understanding| pages=38–59| year=1995| issue=61}} [https://doi.org/10.1006/cviu.1995.1004]</ref> The shapes are constrained by the PDM ([[point distribution model]]) [[Statistical Shape Model]] to vary only in ways seen in a training set of labelled examples.

The '''Active Shape Model''' is a well known contour model used in image processing. An active shape model is constructed by taking a set of points around the outline of an object. As the object moves so do the corresponding points. The distribution of these points can then be examined by [[statistics|statistical]] techniques such as [[principal components analysis]] to find the dominant modes of variation. A combination of these modes is then used to construct a model of how the [[shape]] of an object changes with time.

The shape of an object is represented by a set of points (controlled by the shape model). The ASM algorithm aims to match the model to a new image.

[[File:Operation Of Shape Model In Active Shape Model.jpg|frame|right|Operation of the shape model]]

Active shape models are often used for biological data such as a walking person, and used as a technique for indentifying an object in a scene.

The ASM works by alternating the following steps:

The Active Shape Model Toolkit is a set of software tools for image interpretation developed by Dr [[Tim Cootes]], a research fellow at the Wolfson Image Analysis Unit of the [[University of Manchester]]. The toolkit provides a variety of two-dimensional image analysis techniques intended for applications in industry, medicine and science. The toolkit is available as an add-on for [[MatLab]].

* Generate a suggested shape by looking in the image around each point for a better position for the point. This is commonly done using what is called a "profile model", which looks for strong edges or uses the [[Mahalanobis distance]] to match a model template for the point.<ref name=Cootes/>

* Conform the suggested shape to the point distribution model, commonly called a "shape model" in this context. The figure to the right shows an example.

The technique has been widely used to analyse images of faces, mechanical assemblies and medical images (in 2D and 3D).

It is closely related to the [[active appearance model]]. It is also known as a "Smart Snakes"<ref name=Cootes/> method, since it is an analog to an [[active contour model]] which would respect explicit shape constraints.

==See also==

* [[Procrustes analysis]]

* [[Point distribution model]]

==References==

<references/>

==External links==

* [http://www.mathworks.com/matlabcentral/fileexchange/26706-active-shape-model-asm Matlab code] open-source ASM implementation.

* [https://web.archive.org/web/20060706184806/http://www.isbe.man.ac.uk/val~bim/asmtk/Models/index.html Description] atof AAMs from Manchester University website.

{{comp-stub}}

* [http://personalpages.manchester.ac.uk/staff/timothy.f.cootes/ Tim Cootes'] home page (one of the original co-inventors of ASMs).

* [http://www.milbo.users.sonic.net/stasm Source code] for ASMs (the "stasm" library).

* [https://code.google.com/p/asmlib-opencv/ ASMlib-OpenCV], An open source C++/OpenCV implementation of ASM.

{{DEFAULTSORT:Active Shape Model}}

[[Category:Computer vision]]