Guidelines for Creating a Rule-Based Knowledge Learning System and Their Application to a Chinese Business Card Layout Analysis

Rule selection has long been a problem of great challengethat has to be solved when developing a rule-based knowledge learningsystem. Many methods have been proposed to evaluate the eligibility ofa single rule based on some criteria. However, in a knowledge learningsystem there is usually a set of rules. These rules are notindependent, but interactive. They tend to affect each other and form arule-system. In such case, it is no longer reasonable to isolate eachrule from others for evaluation. A best rule according to certaincriterion is not always the best one for the whole system. Furthermore,the data in the real world from which people want to create their learningsystem are often ill-defined and inconsistent. In this case, thecompleteness and consistency criteria for rule selection are no longeressential. In this paper, some ideas about how to solve therule-selection problem in a systematic way are proposed. Theseideas have been applied in the design of a Chinese business card layout analysis systemand gained a good result on the training data set of 425 images. Theimplementation of the system and the result are presented in this paper.