US 11,694,697 B2
System and method to correct for packet loss in ASR systems
Srinath Cheluvaraja, Carmel, IN (US); Ananth Nagaraja Iyer, Carmel, IN (US); Aravind Ganapathiraju, Hyderabad (IN); and Felix Immanuel Wyss, Zionsville, IN (US)
Filed by GENESYS TELECOMMUNICATIONS LABORATORIES, INC., Daly City, CA (US)
Filed on Jun. 29, 2020, as Appl. No. 16/915,160.
Application 16/915,160 is a division of application No. 16/186,851, filed on Nov. 12, 2018, granted, now 10,789,962.
Prior Publication US 2020/0335110 A1, Oct. 22, 2020
Int. Cl. G10L 19/005 (2013.01); G10L 15/08 (2006.01); G10L 15/14 (2006.01); G10L 15/20 (2006.01); G10L 19/00 (2013.01); G10L 15/02 (2006.01); G10L 25/18 (2013.01); G10L 25/21 (2013.01)
CPC G10L 19/005 (2013.01) [G10L 15/08 (2013.01); G10L 15/14 (2013.01); G10L 15/142 (2013.01); G10L 15/20 (2013.01); G10L 15/02 (2013.01); G10L 25/18 (2013.01); G10L 25/21 (2013.01); G10L 2015/025 (2013.01); G10L 2019/0012 (2013.01)] 18 Claims
OG exemplary drawing
 
1. A method to correct for packet loss in an audio signal in automatic speech recognition systems comprising rescoring of phoneme probability calculations for a section of the audio signal, comprising the steps of:
accumulating probabilities for a given word in a lexicon using the highest scoring tokens for every frame of the audio signal;
determining when the probabilities for the given word are statistically significant;
reporting matches when the probabilities are statistically significant;
normalizing values of probabilities that are not statistically significant by replacing values for section of the audio signal comprising packet loss; and
rescoring the probabilities for the given word.