BDBComp
Parceria:
SBC
Estudo e proposta de adaptação do algoritmo SIFT em relação ao problema de iluminação em imagens

W. R. S. MachadoA. H. F. LouroA. GonzagaM. Boaventura

In recent years, two promising approaches for salient regions detection in images have received special attention from the scientific community: SIFT (Scale Invariant Feature Transform) and SURF (Speeded Up Robust Features). Based in recent studies that describe some limitations of the SIFT, this work has as objective to prove the deficiency of the algorithm in relation to the illumination problem and to pose an adaptation in original algorithm through an innovative approach that uses a process of not homogeneous suavization to improve the performance of the algorithm SIFT.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2008/0044.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