diff options
| author | epapoutsellis <epapoutsellis@gmail.com> | 2019-04-16 19:05:10 +0100 |
|---|---|---|
| committer | epapoutsellis <epapoutsellis@gmail.com> | 2019-04-16 19:05:10 +0100 |
| commit | 48cffabc0879f9e61d932b654e497382ebfaa995 (patch) | |
| tree | 70c548bd743798c95e277cdbbb7089b464e8e35a | |
| parent | e23a45a61a8f185efe564088daea45de714d94ac (diff) | |
| download | framework-48cffabc0879f9e61d932b654e497382ebfaa995.tar.gz framework-48cffabc0879f9e61d932b654e497382ebfaa995.tar.bz2 framework-48cffabc0879f9e61d932b654e497382ebfaa995.tar.xz framework-48cffabc0879f9e61d932b654e497382ebfaa995.zip | |
fix KL methods
| -rw-r--r-- | Wrappers/Python/ccpi/optimisation/functions/KullbackLeibler.py | 13 |
1 files changed, 10 insertions, 3 deletions
diff --git a/Wrappers/Python/ccpi/optimisation/functions/KullbackLeibler.py b/Wrappers/Python/ccpi/optimisation/functions/KullbackLeibler.py index ae25bdb..40dddd7 100644 --- a/Wrappers/Python/ccpi/optimisation/functions/KullbackLeibler.py +++ b/Wrappers/Python/ccpi/optimisation/functions/KullbackLeibler.py @@ -20,6 +20,7 @@ import numpy from ccpi.optimisation.functions import Function from ccpi.optimisation.functions.ScaledFunction import ScaledFunction +from ccpi.framework import ImageData class KullbackLeibler(Function): @@ -39,14 +40,17 @@ class KullbackLeibler(Function): # TODO check - self.sum_value = self.b + self.bnoise + self.sum_value = x + self.bnoise if (self.sum_value.as_array()<0).any(): self.sum_value = numpy.inf if self.sum_value==numpy.inf: return numpy.inf else: - return numpy.sum( x.as_array() - self.b.as_array() * numpy.log(self.sum_value.as_array())) + tmp = self.sum_value.as_array() + return (x - self.b * ImageData( numpy.log(tmp))).sum() + +# return numpy.sum( x.as_array() - self.b.as_array() * numpy.log(self.sum_value.as_array())) def gradient(self, x, out=None): @@ -60,7 +64,10 @@ class KullbackLeibler(Function): def convex_conjugate(self, x): - return self.b * ( numpy.log(self.b/(1-x)) - 1 ) - self.bnoise * (x - 1) + tmp = self.b.as_array()/( 1 - x.as_array() ) + + return (self.b * ( ImageData( numpy.log(tmp) ) - 1 ) - self.bnoise * (x - 1)).sum() +# return self.b * ( ImageData(numpy.log(self.b/(1-x)) - 1 )) - self.bnoise * (x - 1) def proximal(self, x, tau, out=None): |