ISCA Archive Eurospeech 2001
ISCA Archive Eurospeech 2001

Recognition performance of the siemens front-end with and without frame dropping on the Aurora 2 database

Bernt Andrassy, Damjan Vlaj, Christophe Beaugeant

Following the objective of the Eurospeech special event, 'Noise Robust Recognition', the recognition results of a noise robust front-end, developed by Siemens, on the Aurora 2 database [1] are presented in this paper. The front-end was tested with and without a frame dropping algorithm. It is shown that the front-end improves the recognition results in high mismatch between training and testing by 43.90% over the reference front-end and works particularily well in conditions with high noise. Furthermore it is shown that the frame dropping mainly increases the performance of the front-end.