Title :
A neural network for motion computation
Author_Institution :
LETI-DSYS, CEA-CENG, Grenoble
Abstract :
Summary form only given. The author discusses a neural network dedicated to the problem of motion computation, which requires a large amount of time on conventional computers. The initial estimation derived from both optical flow calculation and prediction-is the input of an anisotropic diffusion network, which regularizes the solution and preserves the object boundaries as well
Keywords :
filtering and prediction theory; neural nets; optical information processing; anisotropic diffusion network; motion computation; neural network; object boundaries; optical flow calculation; prediction; Anisotropic magnetoresistance; Computational modeling; Computer networks; Geometrical optics; Image motion analysis; Neural networks; Optical computing; Optical fiber networks;
Conference_Titel :
Neural Networks, 1991., IJCNN-91-Seattle International Joint Conference on
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-0164-1
DOI :
10.1109/IJCNN.1991.155542