A Study for fusion with Pattern Recognition and Artificial Intelligence in the hierarchical structures

在研究中，图像、语音等各种模式的理解已经发展到基于基本人工智能的层次结构中，各种模式的先验知识可以在精细的组件级（点、线、网格）上进行处理。并实现了自顶向下处理（人工智能方法）和自底向上处理（用于层次结构层）的不确定性约简范式。具体地，我们针对手绘图像和彩色图像进行了研究。提出了新的知识模型、算法、知识结构，利用自顶向下的知识，用户交互，并在多媒体网络上进行了连接主义模型、智能归档等实验。

In the researches, The understanding of various kinds of patterns, such as images, spoken words, had been developed based on fundamental artificial intelligence in the hierarchical structures.A priori knowledge of the many kinds of patterns on the pattern processing that can be done at the fine components level (points, lines, mesh). And uncertainty reduction paradigm is implemented for a top-down processing ( methods of the artificial intelligence ) with a bottom-up processing ( use to the hierarchical structures level ).In conconcretely, We are researched for a hand sketched image and color images. We propose new knowledges for the model, algorithms, knowledge structures ; by use the top-down knowledges, user interaction.In a methodicaly, We operate many kinds of experiments such as connectionist-model, intelligent filing by use the multi-media network.