BDBComp
Parceria:
SBC
RECUPERAÇÃO DE IMAGENS DE FACES HUMANAS BASEADA EM CONTEÚDO ATRAVÉS DE WAVELETS

Marcelo Franceschi de BianchiRodrigo Capobianco GuidoThais Lorasqui ScarpaPaula PadovanVladisa S. RamosReinaldo MasukoWellington F. Bastos

This work describes an efficient algorithm for content-based image retrieval based on discrete wavelet transform (DWT) filter banks and Euclidian distance operator, an ordinary criterion for distance measurement. The former is used to turn the query image into a signature-vector, a compressed frame which holds the main features of the original data, and the latter is used to calculate the lowest distance between the query signature vector, and all the signature-vectors stored in the data base, each on related to a particular image. Interestingly and according to the tests, we have found that, for one particular query, the worse the frequency response of the filter bank used for compression is, the better the classification is 96,73% being the best accuracy we have reached. The system´s input consists of a query image and its output corresponds to the most similar image found in the data-bank, according to the distance criterion adopted.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2005/0030.pdf

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato: bdbcomp@lbd.dcc.ufmg.br
     Mantida por:
LBD