In this paper, a new feature for text verification is proposed. The difficulties for the selection of features for text verification (FTV) are first discussed, followed by two principles for the FTV: the FTV should minimize the influence of backgrounds, and it should also be expressive enough for all the texts varied in structures prominently. In this paper, we exploit different block partition methods and introduce two widely used features: the gray scale contrast (GSC) feature to eliminate the background difference, and the edge orient histogram (EOH) feature to distinguish the structure of texts from that of non-texts. A texture classifier can be got by SVM training of pre-labeled data. The candidate text lines can be verified by this classifier. Experimental results show that our feature performs well.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.