Dominant Features Extraction from an image using genetic algorithm

I am looking forward to extracting dominant features from a color image using genetic algorithm to feed it into neural networks for detection of object. I got basic understanding about genetic algorithm but chromosome encoding and fitness function are the two things I am confused about! If I am willing to extract dominant Hue information from a color image using genetic algorithm, how can I do it? how can i encode the chromosomes and define a fitness function? How do I initialize the population for this? The entire image pixels are possible solution and do they need to be included in the initial population? Including an entire image will slow down the process? Please enlighten me about this confusion! Thanks in advance!

Answer 1

[S. Muhammad Hossein Mousavi]

This code of mine is for ACO feature extraction:

https://www.mathworks.com/matlabcentral/fileexchange/104385-aco-image-feature-extraction