This thesis addresses the figure-ground segmentation problem in the context of complex systems for automatic object recognition. Firstly the problem of image segmentation in general terms is introduced, followed by a discussion about its importance for online and interactive acquisition of visual representations. Secondly a machine learning approach using artificial neural networks is presented. This approach on the basis of Generalized Learning Vector Quantization is investigated in challenging scenarios such as the real-time figure-ground segmentation of complex shaped objects under continuously changing environment conditions. The ability to fulfill these requirements characterize the novelty of the approach compared to state-of-the-art methods. Finally the proposed technique is extended in several aspects, which yields a framework for object segmentation that is applicable to improve current systems for visual object learning and recognition.