- author = "Marcos Ivan {Quintana Hernandez}",
- title = "Genetic programming applied to morphological image processing",
- school = "University of Birmingham",
- year = "2004",
- address = "UK",
- month = oct,
- keywords = "genetic algorithms, genetic programming, Mathematical Morphology, Evolutionary Computation, Sub-Machine-Code GP, SMCGP",
-
URL = "
http://etheses.bham.ac.uk//id/eprint/9936/1/QuintanaHernandez2005PhD.pdf",
-
URL = "
http://etheses.bham.ac.uk//id/eprint/9936/",
- size = "156 pages",
- abstract = "This thesis presents three approaches to the automatic design of algorithms for the processing of binary images based on the Genetic Programming (GP) paradigm. In the first approach the algorithms are designed using the basic Mathematical Morphology (MM) operators, i.e. erosion and dilation, with a variety of Structuring Elements (SEs). GP is used to design algorithms to convert a binary image into another containing just a particular characteristic of interest. In the study we have tested two similarity fitness functions, training sets with different numbers of elements and different sizes of the training images over three different objectives. The results of the first approach showed some success in the evolution of MM algorithms but also identified problems with the amount of computational resources the method required. The second approach uses Sub-Machine-Code GP (SMCGP) and bitwise operators as an attempt to speed-up the evolution of the algorithms and to make them both feasible and effective. The SMCGP approach was successful in the speeding up of the computation but it was not successful in improving the quality of the obtained algorithms. The third approach presents the combination of logical and morphological operators in an attempt to improve the quality of the automatically designed algorithms. The results obtained provide empirical evidence showing that the evolution of high quality MM algorithms using GP is possible and that this technique has a broad potential that should be explored further. This thesis includes an analysis of the potential of GP and other Machine Learning techniques for solving the general problem of Signal Understanding by means of exploring Mathematical Morphology.",
-
notes = "Also known as \cite{bham_theses9936}
Supervisor: Ela Claridge and Riccardo Poli",
