An acoustic segment modeling approach to query-by-example spoken term detection | IEEE Conference Publication | IEEE Xplore