说明：  在草图符号的自适应学习中，不同用户的训练样本数量可能不同，保持在不同样本数量下良好的学习效 果成为需要解决的一个重要问题．提出一种自适应的草图符号识别方法，该方法采用与训练样本个数相关的分类 器组合策略将模板匹配方法和SVM统计分类方法进行了高效组合．它通过利用支持小样本学习的模板匹配方法 和支持大量样本学习的SVM 方法，并同时利用草图符号中的在线信息和离线信息，实现了不同样本个数下自适应 的符号学习和识别．基于该方法，文中设计并实现了支持自适应识别的草图符号组件．最后，利用扩展的PIBG Toolkit开发出原型系统IdeaNote．评估表明，该方法可以在24类草图符号分别使用1到2O个训练样本时具有较 高的识别正确率和较好的时间性能．
(In the sketch symbol adaptive learning, the number of training samples of different users may be different, to keep the number of samples under different good learning effect Fruit become an important issue to be resolved. The draft plan proposed an adaptive character recognition method using the number associated with the classification of training samples Combination strategy will be the template matching method and SVM classification method was efficient statistical combination. It supports a small sample study by using a template matching method And support a large number of samples to learn the SVM method, and sketch symbols while using online information and offline information and achieve a number of different samples of adaptive Learning and recognition of symbols. Based on this method, the paper designed and implemented to support adaptive sketch recognition symbol components. Finally, using the extended PIBG Toolkit developed a prototype system IdeaNote. Is shown th)