Title :
A SOM neural network that reveals continuous displacement fields
Author_Institution :
Dept. of Math. & Comput. Sci., R. Mil. Coll. of Canada, Kingston, Ont., Canada
Abstract :
We present a neural network algorithm, derived from the Kohonen self-organized mapping algorithm, for the solution of the problem of matching points in two pictures representing slightly displaced and distorted images of the same objects. We describe it hereafter in the context of a particular application, namely the matching of the images of marker-particles suspended in a moving fluid, seen in two pictures of them taken a small time interval apart. We illustrate the quality of the solutions it produces with representative results obtained for some test problems; in all cases it is outstandingly efficient
Keywords :
fluid mechanics; image matching; physics computing; self-organising feature maps; Kohonen self-organized mapping algorithm; SOM neural network; continuous displacement fields; displaced images; distorted images; marker-particles; moving fluid; Biomedical imaging; Computer science; Educational institutions; Mathematics; Military computing; Neural networks; Remote sensing; Stereo vision; Testing; Visualization;
Conference_Titel :
Neural Networks Proceedings, 1998. IEEE World Congress on Computational Intelligence. The 1998 IEEE International Joint Conference on
Conference_Location :
Anchorage, AK
Print_ISBN :
0-7803-4859-1
DOI :
10.1109/IJCNN.1998.685884
