Title :
Particle tracking by deformable templates
Author :
Yuille, Alan L. ; Honda, Ko ; Peterson, Carsten
Author_Institution :
Div. of Appl. Sci., Harvard Univ., Cambridge, MA, USA
Abstract :
Describes an approach to particle tracking based on deformable templates. Hough transforms are used to give initial conditions for the templates which then converge using a deterministic annealing algorithm. This template approach is closely related to the elastic net algorithm. The authors demonstrate successful results on both simulated and real data. The authors also show how the Hough transform can be derived from the deformable template algorithm as a special case in the low temperature (or gradient descent limit)
Keywords :
convergence; neural nets; physics computing; position sensitive particle detectors; proportional counters; simulated annealing; transforms; Hough transforms; convergence; deformable templates; deterministic annealing algorithm; elastic net algorithm; gradient descent limit; initial conditions; particle tracking; Aircraft; Annealing; Application software; Computer vision; Magnetic fields; Neural networks; Particle tracking; Physics; Psychology; Temperature;
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.1991.155141