Title:
NETWORK MODEL TRAINING METHOD, IMAGE PROPERTY RECOGNITION METHOD AND APPARATUS, AND ELECTRONIC DEVICE
Document Type and Number:
WIPO Patent Application WO/2021/102655
Kind Code:
A1
Abstract:
A network model training method, an image property recognition method, an image property recognition apparatus (500), and an electronic device (600). The network model training method comprises: obtaining an image sample set (101); constructing a basic model and a target loss function corresponding to the basic model (102); training the basic model according to the image sample set and the loss function until the basic model is converged; and using the converged basic model as a recognition model for recognizing image properties (105).
Inventors:
GAO HONGTAO (CN)
Application Number:
PCT/CN2019/120749
Publication Date:
June 03, 2021
Filing Date:
November 25, 2019
Export Citation:
Assignee:
SHENZHEN HEYTAP TECH CORP LTD (CN)
GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD (CN)
GUANGDONG OPPO MOBILE TELECOMMUNICATIONS CORP LTD (CN)
International Classes:
G06K9/62
Foreign References:
| CN109522942A | 2019-03-26 | |||
| CN109871444A | 2019-06-11 | |||
| CN108596090A | 2018-09-28 | |||
| CN110298266A | 2019-10-01 |
Attorney, Agent or Firm:
ESSEN PATENT&TRADEMARK AGENCY (CN)
Download PDF: