-- In this paper, an efficient approach to segment Persian off-line handwritten text-line into characters is presented. The proposed algorithm first traces the baseline of the inpu...
During the last years, the use of string kernels that compare documents has been shown to achieve good results on text classification problems. In this paper we introduce the appl...
In this paper, we demonstrate that accurate machine translation is possible without the concept of “words,” treating MT as a problem of transformation between character string...
Graham Neubig, Taro Watanabe, Shinsuke Mori, Tatsu...
Two equal length strings, or two equal sized two-dimensional texts, parameterize match (p-match) if there is a one-one mapping (relative to the alphabet) of their characters. Two-...
In this paper, we describe a method for automatically retrieving collocations from large text corpora. This method retrieve collocations in the following stages: 1) extracting str...