Prev: analyse data from ethernet cable using matlab7.0
Next: Change range of values in matrix all to single value
From: Chandrama Dey on 24 Feb 2010 23:50
Hi, I am ph.d final year student on UNSW.I am working with image processing and linear unmixing for flood inundation mapping. what ever coding I have found in file exchange is clustering and PCA based on maximum likelihood method. I have a discriminant function:

gi(x)=1/l Si l-(x-mi)' * 1/Si(x-mi)
where= Si=sample variance-covariance matrix for class i
mi= mean of class i
l.l = determinant of the specified matrix
the probability of gi(x) that a pixel vector x of p elements(spectral bands) is a member of class i is gven by the multivariate normal desity.

I need some valuable suggestions in coding this function to run maximum likelihood classification function for an image (a matrix).
 | 
Pages: 1
Prev: analyse data from ethernet cable using matlab7.0
Next: Change range of values in matrix all to single value