Title:
MODEL TRAINING METHOD AND APPARATUS, STORAGE MEDIUM, AND ELECTRONIC DEVICE
Document Type and Number:
WIPO Patent Application WO/2021/087985
Kind Code:
A1
Abstract:
A model training method and apparatus, a storage medium, and an electronic device, the method comprising: acquiring a sample image set; inputting the sample image set into a deep neural network for training; if a classification image is inputted, calculating a loss value on the basis of a first loss function; and if a target detection image is inputted, calculating a loss value on the basis of the first loss function and a second loss function; and performing backpropagation on the basis of the loss value to update a network parameter until convergence, so as to obtain an image recognition model. The described method can improve the accuracy of image category prediction and target detection by a deep neural network.
Inventors:
GAO HONGTAO (CN)
Application Number:
PCT/CN2019/116710
Publication Date:
May 14, 2021
Filing Date:
November 08, 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:
CN110349147A | 2019-10-18 | |||
CN107134144A | 2017-09-05 | |||
CN110189317A | 2019-08-30 | |||
CN110298266A | 2019-10-01 | |||
CN109522967A | 2019-03-26 | |||
US20190251333A1 | 2019-08-15 |
Attorney, Agent or Firm:
ESSEN PATENT & TRADEMARK AGENCY (CN)
Download PDF: