Edge detection by Genetic Algorithm operator

Aim: obtain a perfect operator which can detect the edges better
database: berkley ground truth edge detection images(input & ideal output images pair)
method:operator masking based image edge detection
technique: genetic algorithm
objective function: mse
constraint:sum of all elements in masking operator is zero(look on sobel edge detection mask properties)
genes:9 elements in the operator
steps:
1.load the input image and ideal, expected output image
2.apply GA algoritm and find a 3*3 operator mask
3.perform edge detection on input image using above operator
4.compare result obtained with ideal expected output using GA fitness function and update the mask
5.repeat step 3 till stopping condition
6.display the result
About:
There are many better techniques available for edge detection than this. This work simply give a working model of masking operator using GA. you can further fine tune the algorithm by optimizing other parameters like threshold, adding NMS, etc..

GA just tell the optimal path(operator values) to reach the target solution(ideal edge/boundary image).