Publication: Identificación de imágenes por análisis de texturas utilizando aprendizaje automático
Loading...
Official URL
Full text at PDC
Publication Date
2019
Advisors (or tutors)
Editors
Journal Title
Journal ISSN
Volume Title
Publisher
Citations
Exportar
Abstract
En la actualidad, el reconocimiento de objetos en imágenes a través del procesamiento digital y el aprendizaje automático es un campo que se encuentra en claro auge dentro del mundo moderno, esto es así hasta el punto de que la utilización de las tecnologías derivadas de esta área de investigación han generado múltiples dilemas éticos en la sociedad actual. Para poder aseverar esta afirmación no tenemos más que rememorar las distintas polémicas causadas, recientemente, por los sistemas de reconocimiento facial.
Dentro del procesamiento digital por aprendizaje automático encontramos el análisis de textura como una de las variantes más prominentes en el organigrama de este extenso campo. Y precisamente este será el tema fundamental de nuestro proyecto, el cual desarrollaremos y profundizaremos a lo largo de estas páginas.
Por ende, nuestro objetivo para este proyecto será implementar un entorno operativo que facilite el desarrollo de diferentes aplicaciones de clasificación de texturas y que permita ensayos con diferentes parámetros y configuraciones. Además, aplicaremos el sistema, a modo de ensayo práctico, sobre el problema de identificación de la neumonía a partir de radiografías con rayos X.
Para llevar a cabo esto, el proyecto se dividirá en varias partes diferenciadas:
Investigación y estudio previo.
Desarrollo de la aplicación en tres módulos funcionales.
Testeo para la búsqueda de posibles errores de implementación y búsqueda de la configuración óptima para el problema de la neumonía.
Resultados y conclusiones.
Nowadays, the recognition of objects in images through digital processing and machine learning is a field that is clearly booming in the modern world, this is so to the point that the use of technologies derived from this area of research has generated multiple ethical dilemmas in today's society. To be able to assert this affirmation we have only to remember the different controversies caused, recently, by the facial recognition systems. Within the digital processing by machine learning we find texture analysis as one of the most prominent variants in the organization chart of this extensive field. And precisely this will be the fundamental theme of our project, which we will develop and deepen throughout this pages. Therefore, our goal for this project will be to implement an operating environment that facilitates the development of different textural classification applications and allows testing with different parameters and configurations. In addition, we will apply the system, as a practical test, on the problem of identifying pneumonia from x-rays. To carry out this, the project will be divided into several differentiated parts: Research and previous study. Development of the application in three functional modules. Testing to search for possible implementation errors and search for the optimal configuration for the pneumonia problem. Results and conclusions
Nowadays, the recognition of objects in images through digital processing and machine learning is a field that is clearly booming in the modern world, this is so to the point that the use of technologies derived from this area of research has generated multiple ethical dilemmas in today's society. To be able to assert this affirmation we have only to remember the different controversies caused, recently, by the facial recognition systems. Within the digital processing by machine learning we find texture analysis as one of the most prominent variants in the organization chart of this extensive field. And precisely this will be the fundamental theme of our project, which we will develop and deepen throughout this pages. Therefore, our goal for this project will be to implement an operating environment that facilitates the development of different textural classification applications and allows testing with different parameters and configurations. In addition, we will apply the system, as a practical test, on the problem of identifying pneumonia from x-rays. To carry out this, the project will be divided into several differentiated parts: Research and previous study. Development of the application in three functional modules. Testing to search for possible implementation errors and search for the optimal configuration for the pneumonia problem. Results and conclusions
Description
Trabajo de Fin de Grado, Universidad Complutense, Facultad de Informática, Departamento de Arquitectura de Computadores y Automática, Curso 2018/2019