Active Shape Model
Active Shape Models (ASM) are an image processing method developed by Cootes and Taylor .
They are an extension of the snakes by Kass, Witkin and Terzopoulos, which is why they are also known as "Smart Snakes".
ASM represent a model as a set of interconnected points. These points describe, for example, the outline of a certain object. An average of the object outlines or shapes is generated from training images. The training images can also be used to determine the way in which the shapes can be changed.
If you have created a model and want to use it to search for a specific object in an image, the average shape is first placed on the image. Then the points of the model are iteratively drawn to edges in the image.
The method was later expanded to " Active Appearance Models " (AAM). AAM not only consider the outlines of an object, but also the average gray value distribution within the object boundaries.
literature
- Tim F. Cootes, Chris J. Taylor: Active Shape Models - "Smart Snakes" . In: David Hogg u. a. (Ed.): BMVC92. Proceedings of the British Machine Vision Conference; 22.-24. September 1992, Leeds . Springer-Verlag, Berlin 1992, ISBN 3-540-19777-X , pp. 266-275.