Publica
Hier finden Sie wissenschaftliche Publikationen aus den FraunhoferInstituten. Derivativeaugmented features as a dynamic model for timeseries
 European Association for Signal Processing EURASIP; Institute of Electrical and Electronics Engineers IEEE: 23rd European Signal Processing Conference, EUSIPCO 2015 : August 31  September 4, 2015, Nice Piscataway, NJ: IEEE, 2015 ISBN: 9780992862633 ISBN: 9780992862640 S.958962 
 European Signal Processing Conference (EUSIPCO) <23, 2015, Nice> 

 Englisch 
 Konferenzbeitrag 
 Fraunhofer FKIE () 
Abstract
In the field of automatic speech recognition (ASR), it is common practice to augment features with timederivatives, which we call derivativeaugmented features (DAF). Although the method is effective for modeling the dynamic behavior of features and produces signiicantly lower classiication error, it violates the assumption of conditional independence of the observations. The traditional approach is to ignore the problem (simply apply the mathematical approach that assumes independence). In this paper, we take an alternative approach in which we still use the same mathematical approach as before, but calculate a correction factor by integrating out the redundant dimensions. This makes it possible to compare and combine a DAF PDF and a nonDAF PDF. We conduct experiments to demonstrate the usefulness of the approach.