Extracción de características para la clasificación de imágenes a través de la reducción de dimensionalidad aplicando Análisis de Componentes Independientes (ICA)

Abstract

This research presented by thesis "Extraction feature´s for images classifications via dimensionality reduction using independent components analysis (ICA)" has included a very interesting study of the management of independent components on computer vision area. Several experiments were done to search the best technique to find similarities between independent components of distinct images, what is sought is to compare an image that belongs a group of test images with each image´s data training group and determine which are more similar each other, doing images pairs in basis of which we calculate accuracy percentages of the distinct experiments.