Created by W.Langdon from gp-bibliography.bib Revision:1.8237
In this thesis, we have developed new methods based on multigene genetic programming (MGGP) as an evolutionary learning technique inspired by natural evolution, allowing, from a learning set, to find good models combination of image quality metrics, simultaneously optimizing two competing objectives: the suitability of models for their correlation with the subjective scores of various types of distortions in the images, and the complexity of their structure. Due to the explosive growth of screen-oriented applications, we have exploited the advantages offered by MGGP to develop a generic method for the objective evaluation of images of screen content. The results of experiments carried out on the large benchmarks, showed the superior performance of the proposed methods compared to advanced measurements, including other recently published fusion approaches.",
merzougui:2017:ISPA not found Feb 2025
Supervisor: Leila Djerou",
Genetic Programming entries for Naima Merzougui