Title:
NEURAL NETWORK TRAINING METHOD AND IMAGE MATCHING METHOD AND DEVICE
Document Type and Number:
WIPO Patent Application WO/2020/119311
Kind Code:
A1
Abstract:
Disclosed in the present application are a neural network training method and an image matching method and device. The neural network training method at least comprises: marking annotation information of a first clothing instance and a second clothing instance, the first clothing instance and the second clothing instance respectively coming from a first clothing image and a second clothing image; in response to a situation where the first clothing instance matches the second clothing instance, pairing the first clothing image and the second clothing image; and training a neural network to be trained on the basis of the paired first and second clothing images.
Inventors:
GE YUYING (CN)
WU LINGYUN (CN)
ZHANG RUIMAO (CN)
LUO PING (CN)
WU LINGYUN (CN)
ZHANG RUIMAO (CN)
LUO PING (CN)
Application Number:
PCT/CN2019/114449
Publication Date:
June 18, 2020
Filing Date:
October 30, 2019
Export Citation:
Assignee:
SHENZHEN SENSETIME TECH CO LTD (CN)
International Classes:
G06N3/08; G06V10/25; G06V10/44; G06V10/774
Foreign References:
| CN109670591A | 2019-04-23 | |||
| CN106504064A | 2017-03-15 | |||
| CN107845092A | 2018-03-27 | |||
| CN108022161A | 2018-05-11 |
Attorney, Agent or Firm:
CHINA PAT INTELLECTUAL PROPERTY OFFICE (CN)
Download PDF: