Login| Sign Up| Help| Contact|

Patent Searching and Data


Title:
ALARM METHOD BASED ON RESIDUAL NETWORK, AND APPARATUS, COMPUTER DEVICE AND STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2021/047190
Kind Code:
A1
Abstract:
Disclosed are an alarm method based on a residual network, and an apparatus, a computer device and a storage medium. The method comprises: acquiring a first image collected by a first camera, inputting the first image into a pedestrian re-recognition model to obtain a pedestrian recognition result; if the pedestrian recognition result is different from a suspect object, acquiring a second image; extracting image information of the first object to obtain a micro-expression recognition result; if the micro-expression recognition result does not belong to a pre-set malicious micro-expression list, collecting voice information to obtain a voiceprint recognition result; if the voiceprint recognition result is not a negative voiceprint, extracting image information of a second object from the second image, and inputting the image information into a pre-set interference degree calculation model, thereby obtaining an interference degree value output by the interference degree calculation model; and if the interference degree value is not within a pre-set interference numerical value interval, executing an alarm operation. Therefore, the alarm accuracy is improved.

Inventors:
YANG LING (CN)
Application Number:
PCT/CN2020/088046
Publication Date:
March 18, 2021
Filing Date:
April 30, 2020
Export Citation:
Click for automatic bibliography generation   Help
Assignee:
ONE CONNECT SMART TECH CO LTD SHENZHEN (CN)
International Classes:
G06K9/62; G06K9/00; G06N3/04
Foreign References:
CN110765850A2020-02-07
CN109977893A2019-07-05
CN107016608A2017-08-04
CN101266704A2008-09-17
AU2018100321A42018-04-26
GB2566762A2019-03-27
US20160005050A12016-01-07
CN109063649A2018-12-21
CN106982426A2017-07-25
Attorney, Agent or Firm:
SHENZHEN MINGRIJINDIAN INTELLECTUAL PROPERTY AGENCY FIRM(GENERAL) (CN)
Download PDF: