Frintrop, S.S.FrintropRome, E.E.Rome2022-03-102022-03-102004https://publica.fraunhofer.de/handle/publica/346451We present a new recognition system motivated by human vision for the detection and recognition of objects in real-world images. A computational attention module finds regions of interest in an image and a classifier searches for objects only in these regions. This enables a significantly faster classification. We show how name plates in an office environment are detected by the visual attention module and reliably recognized by the classifier.en005006629conference paper