Image data preparation for intelligent security systems

In this paper, we present a new complex approach for image data preparation (IDP) that is optimized for robust an object of interest (OOI) detection intended for intelligent (unmanned) security systems. The objectives of the image processing are: improvement of an OOI detection as well as emphasizing the object specific features for high representative descriptors generation. Our IDP method and algorithms are based on sensation model of the human eye-brain interaction. Following this model and existing object detection algorithms the suggested IDP method takes into account the relatively large OOI dimensions and applies contrast enhancement and object edges emphasizing for these objects. The method is based on optimization criteria providing an object low false detection. The algorithms are effectively applied in an intelligent person tracking security systems. It is shown that the developed method can substantially improve object detection.