SOFTMAX_CVIP
softmax_cvip() - applies softmax normalization to set of feature vectors in a matrix.
Contents
SYNTAX
a = softmax_cvip( vector, r_softmax)
Input Parameters include :
- vector - An m by n numerical matrix where m is the number of vectors and n is the length of each row vector.
- r_softmax - The parameter for softmax scaling.
Output Parameter includes :
- a - A matrix with the same size as the input 'vector' , where each column(feature vector) is normalized using softmax scaling method.
DESCRIPTION
This function normalizes the values in a feature vector based on the softmax scaling method. Each column of the vector matrix is scaled by the softmax parameter r_softmax given as input by the user.
REFERENCE
1. Scott E Umbaugh. DIGITAL IMAGE PROCESSING AND ANALYSIS: Applications with MATLAB and CVIPtools, 3rd Edition.
EXAMPLE
% input vector vectors = randn(13,6); % Parameter for softmax scalling r_softmax = 3; % Calling function a = softmax_cvip(vectors,r_softmax )
a =
0.4070 0.4750 0.5725 0.3481 0.4537 0.3539
0.5321 0.5029 0.6035 0.5524 0.5081 0.5621
0.5023 0.4269 0.5648 0.4639 0.3652 0.4676
0.4782 0.5190 0.5595 0.5693 0.5451 0.4942
0.3936 0.5239 0.3672 0.4701 0.4856 0.5009
0.6067 0.4538 0.4083 0.5765 0.5829 0.5734
0.6492 0.5014 0.4811 0.6630 0.6104 0.4810
0.5039 0.4856 0.5612 0.5282 0.4009 0.6118
0.5322 0.6360 0.4043 0.4448 0.4396 0.5396
0.4620 0.3818 0.3926 0.5059 0.5569 0.3802
0.4819 0.6609 0.4876 0.4021 0.4323 0.4050
0.5770 0.5320 0.5257 0.4760 0.4867 0.5360
0.3725 0.3958 0.5754 0.4988 0.6321 0.5981
CREDITS
Author: Norsang Lama, February 2017
Copyright © 2017-2018 Scott
E Umbaugh
For updates visit CVIP Toolbox Website