Abstract
A number of researchers have used ANNs with different techniques to
recognize invariant patterns (Lee and Oldham (1990), Burr (1988),
Khotanzad and Lu (1990), Barnard and Casasent (1991), Widrow, Winter
and Baxter (1987)). These techniques can be broadly classified as
preprocessing the input signal to be invariant to geometrical differences.
In this paper the back propagation (BP) and Gradient Range Based
Heuristic (GRBH) algorithms for ANN training are presented and their
learning rates compared.
Users
Please
log in to take part in the discussion (add own reviews or comments).