System for grasping speech meaning of recording audio data based on keyword spotting, and indexing method and method thereof using the system

The indexing method and speech information identifying method using a key word extracted based on information identifying the ignition system, and a system for recording the voice data is started. Indexing unit receives input audio data and perform speech recognition of the phoneme based on a frame-by-frame basis to form a phoneme lattice, and the divided indexing information for a frame of time consisting of a plurality of frames of the system - the said partition indexing information, It includes a phoneme lattice formed by each frame of the time-generate, and stores in the database index. Search unit to search through precise acoustic analysis of the phoneme to the key words that are input by the user by search terms match a phoneme column that matches the search term by comparing the phoneme based on the divided index information stored in the index database and match column to find the voice portions, identifying unit outputs to the user to be identified by the main control over a search result searched by the search unit to identify an utterance of said speech information data.