Name
Abstract | ||
|---|---|---|
In this paper, we investigate the polarity information toimprove the active contour model proposed by Chunming etal. [12]. Unlike the traditional level set formulations, thevariational level set formulation proposed by [12] forcesthe level set function to be close to a signed distance function,and therefore completely eliminates the need of the reinitializationprocedure and speeds up the curve evolution.However, like the majority of classical active contour models,the model proposed by [12] relies on a gradient basedstopping function, depending on the image gradient, to stopthe curve evolution. Consequently, using gradient informationfor noisy and textured images, the evolving curve maypass through or stop far from the salient object boundaries.Moreover, in this case, the isotropic smoothing Gaussianhas to be strong, which will smooth the edges too. For thesereasons, we propose the use of a polarity based stoppingfunction. In fact, comparatively to the gradient information,the polarity information accurately distinguishes theboundaries or edges of the salient objects. Hence, Combiningthe polarity information with the active contour modelof [12] we obtain a fast and efficient active contour modelfor salient object detection. Experiments are performed onseveral images to show the advantage of the polarity basedactive contour. |
Year | DOI | Venue |
|---|---|---|
2009 | 10.1109/CRV.2009.19 | Kelowna, BC |
Keywords | Field | DocType |
polarity information,curve evolution,classical active contour model,efficient active contour modelfor,polarity basedactive contour,active contour model,polarity information toimprove,salient object detection,active contour modelof,combiningthe polarity information,gradient basedstopping function,level set,functional dependency,computer vision,gaussian processes,image segmentation,active contour,solid modeling,edge detection,pixel,image texture,data mining,noise measurement,distance function | Active contour model,Computer vision,Object detection,Image gradient,Pattern recognition,Signed distance function,Image texture,Computer science,Edge detection,Level set,Smoothing,Artificial intelligence | Conference |
ISBN | Citations | PageRank |
978-0-7695-3651-4 | 3 | 0.42 |
References | Authors | |
10 | 3 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Riadh Ksantini | 1 | 82 | 15.39 |
| Farnaz Shariat | 2 | 3 | 0.42 |
| Boubakeur Boufama | 3 | 162 | 22.02 |