A managed approach to computer vision

  • Authors:

  • Zoran Kalafatić;Hrvoje Sertić

  • Affiliations:

  • Department of Electronics, Microelectronics, Computer and Intelligent Systems, Faculty of Electrical Engineering and Computing, Zagreb, Croatia;Research & Development Center, Ericsson Nikola Tesla, Zagreb, Croatia

  • Venue:
  • ICCOMP'06 Proceedings of the 10th WSEAS international conference on Computers
  • Year:
  • 2006

Quantified Score

Hi-index 0.00

Visualization

Abstract

Demanding tasks of computer vision systems require efficient design and effective use of available hardware components. Until recently, most of the computer vision systems were implemented using procedural languages and low-level programming. Although performance efficient, the procedural approach has significant drawbacks in terms of difficult extensibility and complicated implementation of generic solutions. This work describes architectural and implementation issues of a component realized to be a foundation of generic solutions in the field of computer vision. Major architectural decisions, from the definition of main component subsystems, through the usage of design patterns to proper interface definition are explained on actual examples from the component realization. The described implementation practices explain how to efficiently utilize a modern component based platform and a managed programming language. The performance analysis of the designed and implemented system clearly demonstrate the advantages of component based approach for realization of generic computer vision systems, capable of being adopted to various kinds of applications.