Research On Speech Recognition Technology In Field Annotation System

Language communication is the most natural and convenient way of communication, while speech interaction is one of the most direct interaction of human-computer interaction field. The man-machine interactive way is still only use graphics display, keyboard and multi touch during work of field annotation, that makes lots of tedious work about attribute editor. The speech recognition technology is applied to field annotation system based on the research status of intelligent speech interactive technology. The purpose is that make the annotation system understand the human language and speech recognition technology assist us to work. As a result, provide a more convenient, faster, more efficient interactive way for the attribute editor of feature in the field annotation system.This article launches the research on the application of speech recognition technology about attribute editor in the field annotation system. The main research contents include:1. This paper chooses the Microsoft SAPI as development tool based on existing theory and technology of speech recognition. By analyzing its structure and content, summarize the key problems of speech recognition technology which is applied to field annotation system: the development of speech recognition, the rules of grammar definition and the semantic modeling.2. We explore the environment configuration, the core interfaces and application requirements of SAPI development tool, design isolated word speech recognition program development process and develops speech recognition program. With testing the correct identification of speech recognition program under different modes, we select the command and control mode to develop speech recognition program.3. Studying language features and dialogue mode of attribute edit in field annotation system, construct vocabulary relate to attribute edit and sum up grammar structure. We compile and test the grammar rules the custom for attribute editor.4. To construct a semantic model based on grammar rules of attribute, we studied the understanding of Chinese language in characteristics area. By researching method to extract keywords in the SAPI syntax rules, the audio information is mapped to the corresponding functions, according to the semantic model. Then we can edit the feature attribute by audio information.In order to realize the function that edit the attribute editor of feature in the field annotation system by audio information, we choose the Microsoft SAPI as the development tool, construct vocabulary relate to attribute edit, sum up grammar structure. Therefore, it makes the filed annotation system have a certain artificial intelligence.