Pesquisa de Clip Arts Combinando Imagens Raster e Vectoriais
Dec 2012
There is a growing development of search and storage systems for images today. One approach adopted in these systems is content-based image retrieval (CBIR). In the scope of these applications there are users who wish to use clip-art images to include in their presentations and documents. There are many of these images in databases on Internet sites or sold in optical media. Search and retrieval of images in these databases make the user go though a number of large lists manually or use searching methods based on text, which are somewhat inefficient. Those clip art databases are made of vector images or raster images. There are technologies for searching and retrieving for both clip art image types, raster and vectorial, however research has been done separately without taking advantages of both research fields to the problem of recovery and explo- ration of clip art collections. The goal of this thesis is to implement a search engine to retrieve clip art images in databases which include vector and raster images. The solution requires the implementation of a raster to vector image converter, the extraction of features from raster and vector images and the evaluation of the clip art retrieval system.
