Intelligent and Adaptive Decision Forests for Perpetual Knowledge Assimilation

Army perceives precision guidance and battlefield awareness as crucial technologies for the future and Automatic Target Recognition (ATR) forms an important component of these technologies. Specifically, tactical target recognition in air-to-ground infrared imagery will increase the lethality of precision guided missiles and reduce fratricide. However, recognition is difficult when the variations in atmospheric conditions and background cause the clutter/target models to behave significantly different from their statistical patterns. Thus, there is a need to constantly update and modify the representative knowledge base with information from new target and clutter samples while retaining non-redundant information. Based on advances in machine learning and artificial intelligence and their sophisticated applications to intelligent imaging solutions, we propose an Incremental Knowledge Assimilation system to optimally incorporate new and relevant data samples into the knowledge base (adaptive bagging based decision forest classifier) in an incremental fashion with minimal memory and computation requirements. Since the goal of the system is to learn perpetually, it lays higher stress on learning the patterns the each mode of the distribution (or each class) rather than focusing merely on the separation of classes. We present preliminary performance results to validate our methodology.