Avoiding explosive search in automatic selection of simplest pattern codes

作者:

Highlights:

摘要

According to the coding theory of pattern perception, the preferred organization of a pattern is reflected by the simplest code that represents this pattern. The number of codes, out of which the simplest one has to be selected, grows exponentially with the complexity of the pattern. So in computer simulation it would not be wise to generate all codes and then select the simplest one. This would take a lot of computing time. The present study proposes a procedure which avoids this explosion of code generation and yet obtains the simplest code. The central part of this procedure consists of translating the search for a simplest code to a shortest route problem.

论文关键词:Structural information,Coding theory of pattern perception,Hierarchical representation of linear codes,Minimization,Automatic code reduction,Shortest route problem

论文评审过程:Received 17 January 1985, Revised 25 July 1985, Available online 19 May 2003.

论文官网地址:https://doi.org/10.1016/0031-3203(86)90022-1