30-04-2014, 12:27 PM
Differential Evolution for Color Image Segmentation with Histogram and Homogeneity Histogram Difference
Evolution for Color Image Segmentation.pptx (Size: 1.59 MB / Downloads: 13)
Overview of DE Algorithm
Introduced by Storn and Price in 1995.
DE uses parameter vectors as individuals in a population.
The key element distinguishing DE from other population-based techniques is differential mutation operator.
DE Initialization
First, all parameter vectors in a population are randomly initialized and evaluated using the fitness function
The initial NP D-dimensional parameter vectors is:
xji,G where i = 1, 2, ..., NP and
j = 1, 2, ..., D
NP - is number of population vectors
D - is dimension
G - is generation
DE Crossover Operator
The mutated vector’s parameters are then mixed with the parameters of another predetermined vector, the target vector, to yield the so-called trial vector.
Choosing a subgroup of parameters, j (or a set of crossover points) for mutation is similar to a process known as crossover in genetic algorithms or evolution strategies.
Future Plan
Previously, I have done work for gray images.
In future I want to extend my work for video image segmentation using these optimization techniques.