【摘要】 为增强机器自适应筛选能力和减少人工干预并最终实现机器自动认知的目标,通过对贝叶斯网的改进,提出了广义贝叶斯网。在其基础上结合图像模式识别理论,定义了图像知识的集合运算、新知识加入等运算以及基于广义贝叶斯网的图像模式识别框架的知识映射变换、表达、存储的统一的抽象结构。在理论上找到了一种图像模式识别框架中知识表达的统一方法,而且对提高图像模式识别过程的机器参与度具有实际的指导意义。更多还原

【Abstract】 To enhance adaptive filtering capacity of the machine,reducing manual intervention and the eventual realization of the machine auto recognize goal,through the Bayesian network improvements,generalized Bayesian network is proposed.In combination of image pattern recognition,the set operation and new knowledge joining operation are defined,a unified abstract structure for image knowledge mapping transformation,expression,storage is defined.This framework not only finds a unified image knowledge representation approach,but also has practical significance of improving the machine involvement in image pattern recognition.更多还原