US 11,721,329 B2
Method, system and apparatus for multilingual and multimodal keyword search in a mixlingual speech corpus
Arun Kumar, New Delhi (IN); and Abhimanyu Popli, New Delhi (IN)
Assigned to Indian Institute of Technology, Delhi, New Delhi (IN); and Centre for Development of Telematics, New Delhi (IN)
Appl. No. 16/646,315
Filed by Indian Institute of Technology, Delhi, New Delhi (IN); and Centre for Development of Telematics, New Delhi (IN)
PCT Filed Sep. 10, 2018, PCT No. PCT/IB2018/056871
§ 371(c)(1), (2) Date Mar. 11, 2020,
PCT Pub. No. WO2019/049089, PCT Pub. Date Mar. 14, 2019.
Claims priority of application No. 201711032030 (IN), filed on Sep. 11, 2017.
Prior Publication US 2020/0273449 A1, Aug. 27, 2020
Int. Cl. G10L 15/187 (2013.01); G06F 16/33 (2019.01); G06F 16/35 (2019.01); G10L 15/02 (2006.01); G10L 15/22 (2006.01); G10L 15/183 (2013.01); G06F 16/60 (2019.01); G06F 16/2452 (2019.01); G06F 16/635 (2019.01); G10L 15/08 (2006.01)
CPC G10L 15/187 (2013.01) [G06F 16/24522 (2019.01); G06F 16/3337 (2019.01); G06F 16/3344 (2019.01); G06F 16/353 (2019.01); G06F 16/60 (2019.01); G06F 16/635 (2019.01); G10L 15/02 (2013.01); G10L 15/08 (2013.01); G10L 15/183 (2013.01); G10L 15/22 (2013.01); G10L 2015/025 (2013.01); G10L 2015/088 (2013.01)] 3 Claims
OG exemplary drawing
 
1. A method for performing at least a multilingual keyword data search by using an electronic device, the method comprising steps of:
inputting said multilingual keyword data, wherein said multilingual keyword data includes a plurality of multilingual text keywords;
converting said multilingual text keywords to obtain a plurality of articulatory classes and subclasses information, wherein said electronic device includes a multilingual speech-based storage system having a plurality of records of information associated with said articulatory classes and subclasses,
wherein said electronic device is pre-trained on multilingual speech to obtain said plurality of records of information associated with said articulatory classes and subclasses in multilingual speech; and
wherein said step of converting comprises:
converting said multilingual text keywords of the said multilingual keyword data into a sequence of multilingual phonemes; thereafter converting said sequence of multilingual phonemes into a sequence of articulatory classes and subclasses; and
further converting said sequence of articulatory classes and subclasses into said plurality of articulatory classes and subclasses information in a format that can be matched with the said plurality of records;
matching said plurality of articulatory classes and subclasses information with said plurality of records to obtain a result.