Figure 1.
A speech-processing workflow compares three system transcripts, aligns them, applies voting and selects the best-scoring transcript.The process begins with an audio waveform. The same audio input is sent to System 1, System 2 and System 3. Each system produces Transcript 1, Transcript 2 and Transcript 3, respectively. The three transcripts enter an alignment module, which aligns their corresponding content. The aligned outputs pass to a voting module. The voting result produces the best-scoring transcript.

Overview of ROVER

This Feature Is Available To Subscribers Only

or Create an Account

Close Modal
Close Modal