Abstract
This paper describes a novel design environment for cognitive systems tailored to the needs of flexibility, speed and transparence demanded in numerous application domains as e.g., automated visual inspection for quality control, mechatronics, medical applications, and other cognitive tasks. Our QuickCog design environment combines methods from statistical pattern recognition, neural networks, image and signal processing, and exploratory data analysis. The main features of the design environment are visual programming, interactive sample set compilation and sample set oriented processing, automatic feature selection, automatic processing method selection and parameter optimisation, feature space visualisation, and powerful self-configuring classifiers. QuickCog is available on PC/Windows?95 and will be commercialised in August?98.