Man-Machine Voice Communication

摘要: Speech is our most natural way of communicating. Effective integration speech into man-machine communication depends on the nature user interface and application. This justification for grouping together topics voice-enabled interactivity engineering in this chapter. A can be defined as mediator between users machines. It a system that handles entire process responsible provision machine “knowledge”, functionality, available information. must carried out (1) compatible with user’s channels; (2) translate actions (user input) form (instructions/commands) understandable by [1]. For approximately 25 years, number international research development projects have working to adapt interaction needs human beings [2]. However, considerable progress field has only been achieved last 10 notably automatic video-signal processing. While we like think communication, it misleading translates facile building interfaces will provide (e.g., computer or robot). Generally, not means communciation machines [3]. The naturalness should seen reflecting experiences using voice modality communicate other humans. These people share great deal knowledge speaker something cannot said current computers, even if there are ongoing efforts broad contextual knowledge. take granted goes away when listener does understand meaning what say. vision 1990s was one ubiquitous, low-cost, easy-to-use computation everyone. Presently, key technologies rapidly evolving fulfill Computers now ability talk, listen perhaps understand. processing covers range activities future goal enabling skills. rises from confluence low-cost improved algorithms stimulated wide uses technology across spectrum information Advances human-language offer promise nearly universal access on-line services. Since almost everyone speaks understands language, spoken language systems allow