I am working on an extremely simple voice recognition project which I plan on implementing DTW for the template matching. Basically, I'll have stored FFT values of a template of some word that I'm trying to match, store incoming data, do FFT, and run the DTW algorithm to see if it matches the template. My question is, and I can't seem to find the answer on any paper I read about DTW, is how do I determine the maximum cost the algorithm can compute and still be labeled as a match? In other words, if the dtw algorithm computes a cost of value X, what criteria do I need to come up with to determine if that is small enough to be a match?


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.