Recognition of Fuzzy Characters on Oracle-Bone Inscriptions

摘要

In this paper, we propose a novel recognition method for fuzzy Oracle-bone inscriptions. Accurate recognition of fuzzy characters is a difficult problem due to their complex shapes. Our strategy is proposed based on statistical analysis of context and method of combining the Hopfield network. First, according to the oracle inscriptions context statistics information, a candidate font library for the fuzzy characters is established, second, the initial recognition work is processed based on oracle-bone's character Hopfield network, last, the OBI component vector is used to calculate the similarity of fuzzy characters. According to the similarity, it can determine the target oracle-bone's character. Experiments show that the method can reach high recognition accuracy for some OBI which include fuzzy characters.