| 摘要 |
PROBLEM TO BE SOLVED: To generate speaker attribute information to specify a speaker from an inputted voice or the like to store this information in a document together with the voice by generating speaker attribute information from the inputted voice and storing attachment position information of the voice, the inputted voice, and speaker attribute information as set information. SOLUTION: A voice is inputted from a microphone input 6-1 and another sound source 6-9 like a recording medium, and the inputted voice is extracted by a voice detection part 6-2. Sound volume adjustment and noise elimination are performed in a preprocessing part 6-3, and features of the voice are detected in a feature detection part 6-4. A standard pattern 6-8 for comparison which is used to recognize a voice is compared with the detected voice, and the voice is recognized by a recognition part 6-5 and is decided in a decision part 6-6. Next, voice attachment position information, voice information, etc., are stored besides a preliminarily stored text to attach voice attachment coordinate data, speaker discrimination information, information obtained by digitizing processing of the voice, speaker attribute information, and terminal discrimination information to the original text. |
