BDBComp
Parceria:
SBC
João Paulo MasieiroFrancisco Assis da SilvaDanillo Roberto PereiraAlmir Olivette ArteroMarco Antonio Piteri

Um Modelo para Composição de Métricas de Desempenho de Redes

This paper presents the algorithm TSRATP(Traffic Sign Recognition using AdaBoost and Transitionbetween Pixels) proposed for traffic sign detection andrecognition using images. Initially, the AdaBoost algorithm isused to detect the plates from a specific training for trafficsignals, without considering all the information printed on theplates, but only the basic information, such as red contours, thepresence or not of diagonal lines represent prohibition, yellow orred background. Different from several studies in the literature,this strategy reduces the amount of different plates needed forAdaBoost training. In addition, the HSV color space was used toeliminate regions incorrectly detected by AdaBoost. The methodproposed in this paper to perform the recognition of traffic signsis based on transitions which occur between adjacent pixelswhich form the signals.

http://www.lbd.dcc.ufmg.br/colecoes/wvc/2015/058.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