Improved Hybrid Approach for Bilingual Language Recognition from Text

In this paper we address the problem of language identification from a written text. This topic belongs to a wider area of text-to-phoneme mapping where the letters of a written text must be translated into their corresponding phonemes. In the context of text-to-phoneme mapping if the words of the written text belong to a single known language, transcription from the written letters to their phonemes can be done directly. In the situations when the language of the input text is not known, usually the first step is to identify the language and based on this information the text-to-phoneme mapping is done. As a consequence the performance of the language detection module can highly influence the accuracy of letter and word translations. In this paper we analyze the performance of a previously published hybrid approach and we introduce an improved method of language identification from written text.