Kirstein Stephan, Wersing Heiko, Körner Edgar
Honda Research Institute Europe GmbH, Carl-Legien-Str. 30, 63073 Offenbach am Main, Germany.
Neural Netw. 2008 Jan;21(1):65-77. doi: 10.1016/j.neunet.2007.10.005. Epub 2007 Nov 23.
We present a biologically motivated architecture for object recognition that is based on a hierarchical feature-detection model in combination with a memory architecture that implements short-term and long-term memory for objects. A particular focus is the functional realization of online and incremental learning for the task of appearance-based object recognition of many complex-shaped objects. We propose some modifications of learning vector quantization algorithms that are especially adapted to the task of incremental learning and capable of dealing with the stability-plasticity dilemma of such learning algorithms. Our technical implementation of the neural architecture is capable of online learning of 50 objects within less than three hours.
我们提出了一种用于目标识别的具有生物学动机的架构，该架构基于分层特征检测模型，并结合了一种为目标实现短期和长期记忆的记忆架构。特别关注的是针对许多复杂形状目标的基于外观的目标识别任务的在线和增量学习的功能实现。我们提出了一些学习向量量化算法的修改，这些修改特别适用于增量学习任务，并且能够处理此类学习算法的稳定性 - 可塑性困境。我们神经架构的技术实现能够在不到三小时的时间内对50个目标进行在线学习。
