Title :
Velocity estimators of visual motion in two spatial dimensions
Author :
Courellis, S.H. ; Marmarelis, V.Z.
Author_Institution :
Univ. of Southern California, Los Angeles, CA, USA
Abstract :
Elementary velocity estimators (EVEs) of visual motion are extended from one spatial dimension to two spatial dimensions. Principles of operation of the 2D-EVEs are described and guidelines for their design are laid out. There are a number of innovations in the way the receptive fields (RFs) of the 2D-EVEs are formed to create units which detect motion in different spatial directions. A brief background on the operation of 1D-EVEs is given, and the different spatial formations that the RFs of 2D-EVEs can assume to achieve motion discrimination in specific spatial directions are discussed. It is shown that detection of motion in different directions is implemented by changing the shape of the input receptive field of the 2D-EVE. Preliminary simulation results from two special cases of motion are presented
Keywords :
motion estimation; 2D estimators; elementary velocity estimator design; receptive fields; visual motion estimation; Biomedical engineering; Biomedical optical imaging; Data mining; Detectors; Encoding; Guidelines; Motion detection; Motion estimation; Neural networks; Shape;
Conference_Titel :
Neural Networks, 1992. IJCNN., International Joint Conference on
Conference_Location :
Baltimore, MD
Print_ISBN :
0-7803-0559-0
DOI :
10.1109/IJCNN.1992.227187
