Diseño e implementación en hardware de un modelo de retina
artificial
Loading...
Official URL
Full text at PDC
Publication date
2015
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citation
Abstract
El presente trabajo se enfoca en el estudio de las posibilidades del software de apoyo a la visión humana, orientado a pacientes con enfermedades de baja visión, es decir, enfermedades que disminuyen la agudeza visual afectando tanto a la precisión lograda en la retina, como a la amplitud del campo visual. En concreto, nos centramos en los efectos de la degeneración macular asociada a la edad (DMAE). En un primer momento, introducimos al lector en el problema de la discapacidad visual en el mundo, y más concretamente, en el subconjunto de los pacientes de DMAE. Atendiendo a la anatomía del ojo, presentamos las características de la dolencia, sus síntomas. Continuamos el capítulo indicando las tecnologías informáticas más utilizadas en el mundo de la visión artificial; para terminar con él, describiendo los algoritmos de realce de imágenes que, consideramos, pueden ayudar a los pacientes de DMAE. Llegados al capítulo 3, introducimos una descripción de nuestro modelo de retina; junto a una breve descripción anatómica del ojo humano. Explicamos cómo opera nuestro modelo de retina, estableciendo, donde proceda, los correspondientes paralelismos con la función del ojo humano; y describiendo cada una de sus cinco etapas. El capítulo 4 versa sobre el entorno de experimentacion donde se ha implementado, tanto del modelo de retina descrito en el capítulo anterior, como de los algoritmos de realce presentados en el capítulo 2; en cada uno de los dos lenguajes de programación en que se ha desarrollado la Retina, a saber, Octave y C. En el capítulo 5 describimos las metodologias usadas y como se ha paralelizado el código del. En concreto, hablamos de la paralelización realizada en OpenMP. Seguidamente presentamos los resultados de la implantación del software en varios dispositivos, observando y contrastando los tiempos de ejecución de los diferentes desarrollos (desarrollos seriales y desarrollos paralelos), en las diferentes máquinas, finalizando así el último capítulo, el sexto; y concluyendo el documento con las referencias bibliográficas que hemos consultado para este fin.
This work focuses on the study of the possibilities of software to support human vision, aimed at patients with diseases of low vision, that is, diseases that decrease visual acuity affecting both the accuracy achieved in the retina, as the breadth of the visual field. Specifically, we focus on the effects of macular degeneration (AMD). At first, we introduce the reader to the problem of visual impairment in the world, and more particularly in the subset of patients with AMD. In response to the anatomy of the eye, we present the characteristics of the disease, its symptoms. Chapter indicating continued information technologies most used in the world of machine vision; to finish with him, describing the image enhancement algorithms that consider, they can help patients with AMD. We arrived to chapter 3, we introduce a description of our model retina; and a brief anatomical description of the human eye. We explain how to operate our model retina, establishing, where appropriate, the corresponding parallels to the function of the human eye; and describing each of its five stages. Chapter 4 deals with the experimentation environment where it is implemented, both retina model described in the previous chapter, as enhancement algorithms presented in Chapter 2; in each of the two programming languages that have been developed Retina, namely Octave and C. In section 5 we describe the methodologies used and as parallelized code. In particular, we discuss OpenMP parallelization performed in. Then we present the results of the implementation of the software on multiple devices, observing and contrasting the execution times of the different developments (developments serial and parallel developments), on different machines, ending the last chapter, the sixth; and finalizing the document with the references that we consulted for this purpose.
This work focuses on the study of the possibilities of software to support human vision, aimed at patients with diseases of low vision, that is, diseases that decrease visual acuity affecting both the accuracy achieved in the retina, as the breadth of the visual field. Specifically, we focus on the effects of macular degeneration (AMD). At first, we introduce the reader to the problem of visual impairment in the world, and more particularly in the subset of patients with AMD. In response to the anatomy of the eye, we present the characteristics of the disease, its symptoms. Chapter indicating continued information technologies most used in the world of machine vision; to finish with him, describing the image enhancement algorithms that consider, they can help patients with AMD. We arrived to chapter 3, we introduce a description of our model retina; and a brief anatomical description of the human eye. We explain how to operate our model retina, establishing, where appropriate, the corresponding parallels to the function of the human eye; and describing each of its five stages. Chapter 4 deals with the experimentation environment where it is implemented, both retina model described in the previous chapter, as enhancement algorithms presented in Chapter 2; in each of the two programming languages that have been developed Retina, namely Octave and C. In section 5 we describe the methodologies used and as parallelized code. In particular, we discuss OpenMP parallelization performed in. Then we present the results of the implementation of the software on multiple devices, observing and contrasting the execution times of the different developments (developments serial and parallel developments), on different machines, ending the last chapter, the sixth; and finalizing the document with the references that we consulted for this purpose.
Description
Proyecto de Sistemas Informáticos (Universidad Complutense, Facultad de Informática, curso 2014/2015)