一种用于图象序列分析的基于规则的证据收集方法

Rule-Based Evidence Accural System for Image Understanding

本文提出一种用于图象序列分析的基于规则的证据收集方法.情景模型由特征关系图表示,物体动态识别是通过计算可信度和不可信度来完成的.

This paper presents a concept for sequen-tially updating information on scene obje-cts.S(?)ne objects are represented by a attri-buted relational graphs.Dynamic recog-nition of objects is accomplished by a beii-ef/dishelief measure.