Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
SPEECH EMOTION RECOGNITION METHOD AND SYSTEM USING FUSION OF CROWD INFORMATION
Document Type and Number:
WIPO Patent Application WO/2022/198923
Kind Code:
A1
Abstract:
A speech emotion recognition method using fusion of crowd information, comprising the steps of: collecting a user speech signal (S1); preprocessing the speech signal to obtain a Mel spectrogram (S2); trimming silence segments before and after the Mel spectrogram (S3); obtaining depth crowd information by means of a crowd classification network (S4); obtaining Mel spectrogram depth information by means of a Mel spectrogram preprocessing network (S5); obtaining fused information by means of an SENet fused feature (S6); and obtaining an emotion recognition result by means of the classification network (S7). According to the method, crowd information features are fused, so that emotion feature extraction is more accurate, and information fusion is performed by means of an SENet channel attention mechanism, so that depth features can be effectively extracted and the overall recognition precision is improved.

Inventors:
LI TAIHAO (CN)
ZHENG SHUKAI (CN)
LIU YULONG (CN)
PEI GUANXIONG (CN)
MA SHIJIE (CN)
Application Number:
PCT/CN2021/115694
Publication Date:
September 29, 2022
Filing Date:
August 31, 2021
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
ZHEJIANG LAB (CN)
International Classes:
G10L25/63; G10L25/24; G10L25/30
Domestic Patent References:
WO2019225801A12019-11-28
Foreign References:
CN112712824A2021-04-27
CN110021308A2019-07-16
CN110491416A2019-11-22
CN109817246A2019-05-28
CN110164476A2019-08-23
CN110852215A2020-02-28
CN108899049A2018-11-27
CN109146066A2019-01-04
CN111292765A2020-06-16
CN108154879A2018-06-12
Other References:
LU GUANMING;YUAN LIANG;YANG WENJUAN;YAN JINGJIE;LI HAIBO: "Speech emotion recognition based on long short-term memory and convolutional neural networks", JOURNAL OF NANJING UNIVERSITY OF POSTS AND TELECOMMUNICATIONS (NATURAL SCIENCE), vol. 38, no. 5, 13 November 2018 (2018-11-13), CN, pages 63 - 69, XP055969218, ISSN: 1673-5439, DOI: 10.14132/j.cnki.1673-5439.2018.05.009
Attorney, Agent or Firm:
HANGZHOU QIUSHI PATENT OFFICE CO., LTD (CN)
Download PDF: