Abstract: The invention relates to a method for controlling data-processing software modules (4, 5, 6, 7), wherein the data to be processed are organized in files, and wherein the processing is performed in a plurality of successive processing sections and the files in one processing section are processed by at least one software module (4, 5, 6, 7). To prevent the propagation and spreading of deviations occurring during the processing in the software modules (4, 5, 6, 7) in subsequent processing sections, information about the data processed are collected in the software modules (4, 5, 6, 7) of the processing sections and are signalized to a control system (8), said control system (8) comparing the information signalized and stopping the processing in a subsequent processing section if the comparison of the information revealed that a predetermined, admissible maximum deviation has been exceeded.
Type:
Grant
Filed:
May 5, 2005
Date of Patent:
October 19, 2010
Assignee:
USB AG
Inventors:
Peter Dunki, Hansbeat Loacker, Markus Wietlisbach
Abstract: The invention is a system and method for automatic acoustic speaker adaptation in an automatic speech recognition assisted transcription system. Partial transcripts of audio files are generated by a transcriptionist. A topic language model is generated from the partial transcripts. The topic language model is interpolated with a general language model. Automatic speech recognition is performed on the audio files by a speech recognition engine using a speaker independent acoustic model and the interpolated language model to generate semi-literal transcripts of the audio files. The semi-literal transcripts are then used with the corresponding audio files to generate a speaker dependent acoustic model in an acoustic adaptation engine.