BDBComp
Parceria:
SBC
Esteno : Uma Abordagem para Detecção Visual de Bankers

Victor F. MartinsAndré R. A. GrégioVitor M. AfonsoPaulo Lício de Geus

Bankers--Internet Banking information stealer programs--usually present windows that mimic legitimate bank sites to lure users into providing sensitive information. In addition, bankers may run on the target operating system in a non-intrusive mode, making the detection and analysis provided by unsupervised, automated analysis systems difficult. In this paper, we propose a solution for the identification of Brazilian bankers. To this end, we leverage three visual analyzers (based on color properties, known logotype presence and textual patterns) that are tuned using a supervised machine learning technique (Random Forest). We tested our approach on over 1,100 unknown binaries´ images, yielding 92.1% of correctly classified samples.

http://www.lbd.dcc.ufmg.br/colecoes/sbseg/2014/0063.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