| CPC G10L 15/26 (2013.01) [G10L 15/22 (2013.01); H04M 1/72448 (2021.01); G10L 15/183 (2013.01); H04M 2201/40 (2013.01)] | 16 Claims |
|
1. A system for identifying spoken language in a telecommunications relay service (“TRS”) wherein automatic speech recognition (“ASR”) engines transcribe and convert peer voices into text packets, and the text packets transcribed from the peer voices are transmitted to a user, the system comprising:
a call serving entity; and
a plurality of ASR groups, each of the ASR groups including an associated ASR engine that recognizes and transcribes speech to a predefined language,
wherein one of the plurality of ASR groups is set as a default ASR group,
wherein the call serving entity is configured to:
establish and release a voice path between the call serving entity and a peer, and a voice path between the call serving entity and the plurality of ASR groups;
establish and release a data path between the call serving entity and the user, and a data path between the call serving entity and the plurality of ASR groups;
receive the peer voices via the voice path and transmit the peer voices to the default ASR group, wherein the peer voices are transcribed and converted into the text packets by the default ASR group;
receive the text packets from the default ASR group;
set, unset, or maintain the default ASR group; and
transmit the text packets to the user via the data path and/or discard the text packets,
wherein the call serving entity is configured to check a confidence level of each text packet received from the default ASR group or from any of the plurality of ASR groups,
wherein any text packet received from the default ASR group is transmitted to the user via the data path and the default ASR group is maintained if the confidence level is higher than a predefined level,
wherein the call serving entity is configured to check a confidence level of each of the text packets received from the default ASR group or from any of the plurality of ASR groups,
wherein any of the text packets received from the default ASR group or the plurality of ASR groups is discarded if the confidence level is lower than a predefined level,
wherein if the confidence level of a first text packet received from the default ASR group is lower than the predefined level, the call serving entity is configured to increase a retry count incrementally, and receive a second text packet from the default ASR group.
|