计算机识别藏文音节构件的方法研究

Method of computer recognition of Tibetan syllable components

藏文音节构件识别是藏文信息处理应当特别关注的一个问题,要实现藏文排序、藏文拉丁转写、藏文文本校对等工作就必须先识别出构成藏文音节的7大构件。针对符合藏文字性组织法构件组合规则的藏文音节,依据藏文字性组织法规定的音节组合规则和组合结构,提出先确定藏文音节中作为核心构件的基字,再依据基字判断出其他构件的算法,结合此算法对藏文中出现的其他特殊音节进行了特殊的构件识别处理。通过测试验证算法的可行性,测试结果表明,该算法能够正确识别符合组合规则和结构的藏文音节,对特殊音节也有较好的识别能力。

Tibetan syllable component recognition should be one of the particular concerns in Tibetan information processing.To achieve the works of Tibetan sorting,Tibetan Latin transliteration and Tibetan text proofreading,the seven components form?ing the Tibetan syllables must be recognized first.Aiming at the Tibetan syllables conforming to the component combinationrules in the Tibetan grammatical work,a method to determine the root as a core component in the Tibetan syllable is proposedin accordance with syllable composition rules and combined structures specified in the Tibetan grammatical work.The algorithmof the other components is judged according to the root.In combination with this algorithm,the specified component recognitionprocessing is conducted for other special syllables occurring in Tibetan.The feasibility of the specified was verified in tests.Thetest results show that the algorithm can correctly recognize the Tibetan syllable conforming to combination rules and structures,and has good recognition capacity for special syllables.