Show simple item record

dc.contributor.advisorOrtiz-Álvarez, Jorge
dc.contributor.authorSilva-Lavalle, Arturo R.
dc.date.accessioned2019-05-14T18:22:52Z
dc.date.available2019-05-14T18:22:52Z
dc.date.issued2006
dc.identifier.urihttps://hdl.handle.net/20.500.11801/2253
dc.description.abstractLas Memorias Asociativas Morfológicas (MAMs) han probado ser un excelente método para identificación de patrones en presencia de ruido tanto dilativo como erosivo; pero por separado. Un método para identificar ruido erosivo y dilativo combinado fue desarrollado en el pasado, al cual se le denominó el método “Kernel”, este requiere de la identificación de un conjunto de vectores (“Kernels”) que son representativos de cada uno de los patrones. Esta tarea de reconocimiento es la debilidad de dicho método debido a que se requiere que la selección de estos “Kernels” resulte ser más una habilidad o arte, que no se ha modelado matemáticamente. El método propuesto, es una alternativa de solución del problema mencionado. Este método hace uso de un Algoritmo Genético para identificar tales vectores “Kernel” procurando conseguir una solución óptima. Dichos “Kernels” óptimos son probados frente a un conjunto de datos. En realidad estos conjuntos de datos son conjuntos de patrones distorsionados aleatoriamente por ruido erosivo y dilativo simultáneamente. Este método y las pruebas correspondientes fueron implementados en MATLAB®.en_US
dc.description.abstractThe Associative Morphological Memories (MAMs) have proved to be an excellent method for identification of patterns in the presence of dilative and erosive noise; but separately. A method to identify erosive and dilative noise simultaneously was developed in the past, to which it was named the “Kernel” method; the algorithm requires identification of a set of vectors (“Kernels”) that are representative of each of the original patterns. This task of recognition is the weakness of the above­mentioned method because it is needed that the selection of this “Kernel” turns out to be more a skill or art, which has not been modeled mathematically. The proposed method is an alternative solution to the mentioned problem. This method makes use of a Genetic Algorithm to identify such “Kernel” vectors trying to obtain an ideal solution. These optimal kernels are tested using a set of data. This set of data is a collection of random noisy patterns containing erosive and dilative noise. This method and respective tests was implemented in MATLAB©.en_US
dc.language.isoSpanishen_US
dc.titleUn método de algoritmos genéticos para optimización de memorias asociativas morfológicasen_US
dc.typeThesisen_US
dc.rights.licenseAll rights reserveden_US
dc.rights.holder(c) 2006 Arturo Rubén Silva-Lavalleen_US
dc.contributor.committeeBorges, José
dc.contributor.committeeRodríguez, Néstor
dc.contributor.representativePortela, Genock
thesis.degree.levelM.S.en_US
thesis.degree.disciplineComputer Engineeringen_US
dc.contributor.collegeCollege of Engineeringen_US
dc.contributor.departmentDepartment of Electrical and Computer Engineeringen_US
dc.description.graduationYear2006en_US


Files in this item

Thumbnail

This item appears in the following Collection(s)

  • Theses & Dissertations
    Items included under this collection are theses, dissertations, and project reports submitted as a requirement for completing a graduate degree at UPR-Mayagüez.

Show simple item record

All rights reserved
Except where otherwise noted, this item's license is described as All Rights Reserved