| ECVision indexed and annotated bibliography of cognitive computer vision publications This bibliography was created by Hilary Buxton and Benoit Gaillard, University of Sussex, as part of ECVision Specific Action 8-1 The complete text version of this BibTeX file is available here: ECVision_bibliography.bib
J. Heikkonen and A. Bulsari
ABSTRACT
Computer vision has long been an important research field in artificial intelligence. This research is motivated by needs for flexible systems capable of performing complex recognition, inspection, assembly and navigation tasks in diverse applications ranging from optical character recognition in postal automation to computer integrated manufacturing in industrial plants. A computer vision system typically consists of consecutive levels. At the lowest level there are raw images consisting as a set of picture points each representing either a color, a brightness, or a range (i.e. depth) value. The middle levels, also known as intermediate levels, typically transform the raw sensory information to a more meaningful form for the given application. The outputs from the uppermost level should finally provide the information needed to solve the problem of interest. Neural networks have been seen as useful tools for different kinds of computer vision applications. Much of the current research in neural networks is centered on individual network models, whereas in complicated engineering tasks such as in computer vision, a system level view of neural networks is more desirable. Individual neural networks are then seen as components in a computer system which alsocontains more traditional image processing techniques, such as filtering and segmentation of images. This kind of use of neural networks leads to a hybrid architecture in which some of the processing modules are based on neural networks. The problem is to then decide what benefits neural networks may provide for a computer vision problem (if any) and what are the most appropriate levels for neural networks and what kinds of neural network models should be used. This above system level view of neural networks can clearly be seen from this special issue devoted to neural networks in computer vision applications. This special issue is based on a few selected papers presented at the International Conference on Engineering Applications of Neural Networks (EANN'95) held in Otaniemi (Espoo/Helsinki, Finland) on 21-23 August, 1995. |
|
| Site generated on Friday, 06 January 2006
|