| 摘要 |
<p>PURPOSE:To determine a collated person surely when calculated distance satisfies a specified condition by calculating the distance between an input pattern and a confirmed pattern and the distance between an input pattern and a suppress reference pattern designated by suppress designation basing on a registered information designating signal. CONSTITUTION:Plural suppress patterns are stored in the suppress reference pattern storing section 90 of a speaker recognizing system, and a confirming reference pattern to be registered is stored in a confirming reference pattern storing section 50 through an input buffer section 40. Further, information that designates a suppress reference pattern resembling to the confirming reference pattern is stored in a suppress designating section 80, and an input pattern by the voice of a person to be collated is held in a buffer 40. The distance between the input pattern and confirming reference pattern, and the distance between the input pattern and suppress reference pattern designated by the designating section 80 are calculated by a pattern matching section 60. When conditions, distance from the reference pattern is less than distance from the suppress reference pattern and the distance from the reference pattern is less than specified threshold value are satisfied, a judging section 110 determines the person himself.</p> |
