Using context knowledge to improve the accuracy of recognition in voice browsing

The accuracy of speech recognition is still a bottleneck to baffle the application of speech technology. We may improve the accuracy of recognition greatly by using context knowledge efficiently. In Web speech browsing, a user´s speech input is usually one element of a finite set. Based on these observations, the paper first defines a kind of similarity between two Chinese text strings, then processes the recognition results of the engine to acquire more accurate results. Experiments show that our approach is mostly efficient.