US 12,254,877 B2
Display device
Eunkyung Jung, Seoul (KR); Hyosung Kim, Seoul (KR); Taejin Park, Seoul (KR); Woojin Choi, Seoul (KR); and Sunho Hwang, Seoul (KR)
Assigned to LG ELECTRONICS INC., Seoul (KR)
Appl. No. 17/298,319
Filed by LG ELECTRONICS INC., Seoul (KR)
PCT Filed Dec. 11, 2018, PCT No. PCT/KR2018/015684
§ 371(c)(1), (2) Date May 28, 2021,
PCT Pub. No. WO2020/122271, PCT Pub. Date Jun. 18, 2020.
Prior Publication US 2022/0028380 A1, Jan. 27, 2022
Int. Cl. G10L 15/22 (2006.01); G10L 15/02 (2006.01); H04N 21/422 (2011.01)
CPC G10L 15/22 (2013.01) [G10L 15/02 (2013.01); G10L 2015/025 (2013.01); G10L 2015/223 (2013.01); H04N 21/422 (2013.01)] 8 Claims
OG exemplary drawing
 
1. A display device comprising:
a wireless communication unit configured to communicate with an external device that receives a voice wake-up word;
a main processor configured to:
receive, from the external device, wake-up word recognition information indicating that a recognition of the voice wake-up word is successful and a speech data of the voice wake-up word through the wireless communication unit,
extract speech feature vectors representing feature information of the voice wake-up word,
convert the speech feature vectors into a combination of phonemes by using a first model,
convert the combination of phonemes into a language text by using a second model,
determine whether the converted language text falls within a recognition range of the normal wake-up word through comparison between registered normal wake-up words previously stored and the converted language text, and
when the converted language text falls within the recognition range of the normal wake-up word, determine the voice wake-up word as the normal wake-up word,
when it is determined that the voice wake-up word is the normal wake-up word, perform a function corresponding to a speech command received from the external device after receiving the speech data of the voice wake-up word,
wherein a number of phoneme combinations and a number of converted language texts are the same,
wherein the converted language texts are divided into the normal wake-up word and an abnormal wake-up word based on the determination whether the converted language text falls within a recognition range of the normal wake-up word,
wherein the abnormal wake-up word is stored in a storage unit to prevent misrecognition, and
wherein the main processor is further configured to inactivate a speech recognition mode of the display device when the abnormal wake-up word is received again, and output a notification informing the reception of the abnormal wake-up word.