Didn't think I'd come across a DTW paper here.

Have you seen anything worth reading in that line of literature which addresses a more practical issue of segment sizing and segment overlap on accuracy?