作者: Mohammad Akbari , Reza Azimi
DOI: 10.1117/12.856302
关键词:
摘要: In this paper we introduce a new approach for information retrieval from Persian document image database without using Optical Character Recognition (OCR).At first an attribute called subword upper contour label is defined then, a pictorial dictionary constructed based on the subwords. By address two issues in document retrieval: keyword spotting and according to similarities. The proposed methods have been evaluated database. results have proved ability of retrieval.