Connected spoken word recognition algorithms by constant time delay DP, O( n) DP and augmented continuous DP matching

Abstract

Dynamic programming (time warping) pattern matching algorithms are proposed for connected spoken word recognition. They are the constant time delay, O( n) DP, and augmented continuous DP matching algorithms. These are based on the same principle as the two level DP, level building DP, and clockwise DP matching methods. The constant time delay DP matching algorithm is an extension of the two level DP. The O( n) DP matching algorithm is computationally more efficient than the conventional methods and also does not require as much memory storage. The amount of computation is the same as that of isolated spoken word recognition by DP matching. Of course, these algorithms give the same results as the two level DP, level building DP, and clockwise DP methods. The augmented continuous DP matching algorithm obtains only a near-optimal solution for the recognition principle based on pattern matching. However, it is computationally more efficient and does not require much memory storage. Further, it could be extended easily to connected word recognition with syntactical constraints.