A simple biologically inspired feature extraction algorithm is proposed for object recognition. First, a set of statistical topographic filters modelling the properties of complex cells in a primary visual cortex (V1) are learned based on enhanced in