歧义图形的机器理解初探

运用突变理论，我们首先对歧义图形的变化规律进行了理论分析，归纳出一类歧义图形折突变模型及其性状。然后在此基础上，具体提出了一种歧义图形理解的计算原理，包括吸引分叉网、理解区分计算以及具体系统的构造。实验结果表明，我们的初步理论是行之有效的，对一般歧义图形的理解研究具有一定的指导意义。

TENTATIVE RESEARCHES ABOUT MACHINE UNDERSTANDING OF AMBIGUOUS PICTURES

By applying the catastrophe theory, we analyse the morphogenesis laws of ambiguous pictures and conclude the catastrophe models from a kind of ambiguous pictures. Then, based on these models, we also provide the computational priciplo of machine understanding for ambiguous, including the attracting-bifurcating not, the understanding-distingushing method and the concrete experiment system Experiment results show that the tentative theory is successful and effective in resolving our problem, and has the guiding significance for tho general researches in ambiguous pictures' understanding.