Active Contours: Snakes
Published Online: 14 DEC 2007
Copyright © 2007 by John Wiley & Sons, Inc.
Wiley Encyclopedia of Computer Science and Engineering
How to Cite
Yilmaz, A. 2007. Active Contours: Snakes. Wiley Encyclopedia of Computer Science and Engineering. .
- Published Online: 14 DEC 2007
The shape of a real-world object can be represented by its outline in the image plane. In computer vision, the outline of the object is referred to as the object contour. A fundamental approach to finding automatically the object contour is the “snakes framework,” which was introduced by the seminal work of Kass, et al. in 1987. For the last two decades, snakes have been used successfully in the context of facial animation, visual speech analysis, traffic monitoring, surveillance, medical imaging (tracking and segmentation of organs), and blue screening in Hollywood movies.
A snake is an elastic model of a continuous and flexible curve that is fitted on the boundary between the object and the rest of the image by analyzing the visual image content. The process of fitting iteratively an initial snake to the object, such that the snake encloses the object tightly, is called “snake evolution.” During its evolution, the snake imposes continuity and smoothness constraints on the evolved contour, which relax the requirement of a noise-free image. In addition to the continuity and smoothness constraints, snakes have the capability to be attracted to certain shape configurations known a priori. The evolution of a snake from one configuration to another in consecutive frames of a video clip attributes a dynamic behavior to the contour and provides object-tracking capabilities. The snake performing object tracking is considered a dynamic contour moving from frame to frame.
- active contour