ABSTRACT -- Speaker identification for Thai language project has been initiated by the National Electronics and Computer Technology Center (NECTEC) since 1999. The first objective is to research and develop a text-dependent closed-set speaker identification system in the office environment. The speaking texts for this system are isolated digit utterances 0-9 and their concatenation. This paper gives an overview of the system, explains the route of the past 1-year research history, and some details of the latest identification system, which achieves the best performance of 92.30% for isolated digit "0" and enhances to 98% for 3-concatenated digit.

KEYWORDS -- Speaker Identification, Text Dependent, Closed Set, Thai Language