Title :
A neural network based recognition of complex two-dimensional objects
Author :
Lauterbach, Bernd ; Besslich, P.W.
Author_Institution :
Bremen Univ., Germany
fDate :
28 Oct-1 Nov 1991
Abstract :
The authors describe the recognition of complex two-dimensional objects using a neural network. The procedure is based on a set of image processing algorithms which produce several feature vectors. These are fed into a hierarchical structured configuration of neural networks. Most of the image processing algorithms can easily be parallelized, so that the implementation on a multiprocessor system is straightforward. The procedure has been implemented on a tree-structured transputer network
Keywords :
image processing; image recognition; neural nets; parallel architectures; transputers; feature vector; hierarchical structured configuration; image processing algorithms; learning algorithm; multilayer perceptions; multiprocessor; neural network; online correction; tree-structured transputer network; two-dimensional objects; Distortion measurement; Fault tolerance; Gravity; Hardware; Image processing; Image segmentation; Length measurement; Neural networks; Quantization; Shape;
Conference_Titel :
Industrial Electronics, Control and Instrumentation, 1991. Proceedings. IECON '91., 1991 International Conference on
Conference_Location :
Kobe
Print_ISBN :
0-87942-688-8
DOI :
10.1109/IECON.1991.238936
