A geometrically inspired measurement of uncertain is proposed in this paper, and this measurement is considered when combining multiple classifiers using D-S evidence theory.

• 本文从几何角度出发，提出了分类器结果的不确定性测量方法，在利用D-S方法合成多个分类器的输出时，将这些不确定性信息也融入计算之中。