Continuous speech recognition on the bases of vector field model for segmentation and feature extraction, and continuous dynamic programming for pattern matching.