Evolvable Image Filter Design with Multi-objectives
Zhiguo Bao, Yi Wang · 2012
In this research, some proposals are raised in evolution process by fixing whole or part of the chromosome representation of the image filter, together with other genetic operators one-point mutation and deterministic selection. It aims to find out better circuit structure by using circuits from previous experiments with best fitness values through mutation and crossover. Therefore, the elite fitness values and Mean Difference per Pixel (MDPP) values should be improved than earlier methods. They are carried out in processes as: (1) chromosome representations of the proposed circuit are defined, either fixed or partially-fixed, (2) multi-objectives are defined mainly to fitness function, (3) performances are evaluated by elite fitness values of the circuit and MDPP values of the filtered images.