两步动态编程识别连续单词的算法实现

Two-level Dynamic Programming Based Algorithm for Connected Word Recognition

传统的DTW算法大多用来识别孤立单词,并已极具实用价值,但是这种方法不适用于连续单词的识别。利用两步动态编程算法解决了连续单词的识别问题。两步动态编程算法首先把每个参考模型与测试串的一部分匹配,然后根据获得两者间最小距离的次序将参考模型个体拼接起来当作辨识解。给出了两步动态编程算法的基本实现方法,并进行了实验验证。

Traditional DTW algorithm is used to recognize isolated word and is proved to be very practical. However, this algorithm can't be used to recognize connected word. This article introduces two- levels dynamic programming to recognize connected word. At the first level the algorithm matches each individual word reference pattern against an arbitrary portion of the test string, the second level of the computation pieces together the individual reference pattern scores to minimize the overall accumulated distance over the entire test string. In the end the article reports an experiment results based on this algorithm.