In this paper we introduce a bio-inspired approach that we apply in the "Bionic Eyeglass Project" which meant to help the everyday life of blind people. The basic ides is to mimic, the nervous system, how it filters out the currently relevant information from the Irrelevant mass - namely realize an attention model. The framework is a complete bottom-up based attention model where the parameters are adjusted via human tests. The principles are firstly the recently discovered ten different mammalian retina channels [3, 4], secondly the saliency map generation and finally how galiency depends on the receptive field types. In the first part of the paper we introduce the theoretical background while the second part contains the empirical results.