Embedded Real-Time Architecture for Level-Set-Based Active Contours
- PDF / 1,166,336 Bytes
- 16 Pages / 600 x 792 pts Page_size
- 77 Downloads / 243 Views
Embedded Real-Time Architecture for Level-Set-Based Active Contours Eva Dejnoˇzkova´ Centre of Mathematical Morphology, School of Mines of Paris, 35 Rue Saint Honor´e, 77305 Fontainebleau Cedex, France Email: [email protected]
´ Petr Dokladal Centre of Mathematical Morphology, School of Mines of Paris, 35 Rue Saint Honor´e, 77305 Fontainebleau Cedex, France Email: [email protected] Received 14 June 2004; Revised 5 April 2005; Recommended for Publication by Luciano da F. Costa Methods described by partial differential equations have gained a considerable interest because of undoubtful advantages such as an easy mathematical description of the underlying physics phenomena, subpixel precision, isotropy, or direct extension to higher dimensions. Though their implementation within the level set framework offers other interesting advantages, their vast industrial deployment on embedded systems is slowed down by their considerable computational effort. This paper exploits the high parallelization potential of the operators from the level set framework and proposes a scalable, asynchronous, multiprocessor platform suitable for system-on-chip solutions. We concentrate on obtaining real-time execution capabilities. The performance is evaluated on a continuous watershed and an object-tracking application based on a simple gradient-based attraction force driving the active countour. The proposed architecture can be realized on commercially available FPGAs. It is built around generalpurpose processor cores, and can run code developed with usual tools. Keywords and phrases: level set, partial differential equations, object tracking, real-time execution, embedded platforms.
1.
INTRODUCTION
The level set was proposed in 1988 in [1] as a simple method to modelize or analyze the motion of a travelling interface. It offers a convenient and stable framework to implement a large variety of methods where images are seen as sets of curves. Since then, its applications have been extended to other image processing fields such as the restoration (filtering or contrast enhancement), segmentation (active contours, watershed) to the form analysis (shortest path, shape-fromshading). See [2] or textbooks [3, 4] for applications and a general overview. From the implementational point of view, the methods can be divided into two groups: (i) filtering-like methods operating on a set of constant-level curves describing the entire image and (ii) methods that act on a single (or several) contour(s), representing one (or several) object(s) present in the image. Below, we reference these algorithms according to their computation scope, the filtering-like methods as globalscope-type and the active contours methods as narrowband type. 1.1. Scope and objectives The objective of this paper is to open the world of hand-held, mobile devices such as PDAs, still picture or movie cameras
or mobile phones to powerful image processing methods from the level set framework. The novelty of this paper resides in the presentation of a reusable architecture capabl
Data Loading...
