Title:
ADVERSARIAL NETWORK ARCHITECTURE OPTIMIZATION METHOD AND SYSTEM, AND IMAGE DESCRIPTION GENERATION METHOD AND SYSTEM
Document Type and Number:
WIPO Patent Application WO/2021/088935
Kind Code:
A1
Abstract:
An adversarial network architecture optimization method and system, an image description generation method and system, an electronic device, and an apparatus. The method comprises the following steps: constructing an adversarial network architecture (S10); and performing, by means of a training image, adversarial optimization training on the adversarial network architecture to obtain an optimized adversarial network architecture (S20). By means of the method, the technical problem of the image description generation methods provided in the related art having a poor generation quality can be solved.
Inventors:
YANG MIN (CN)
LIU JUNHAO (CN)
LI CHENGMING (CN)
JIANG QINGSHAN (CN)
LIU JUNHAO (CN)
LI CHENGMING (CN)
JIANG QINGSHAN (CN)
Application Number:
PCT/CN2020/126829
Publication Date:
May 14, 2021
Filing Date:
November 05, 2020
Export Citation:
Assignee:
SHENZHEN INST OF ADV TECH CAS (CN)
International Classes:
G06N3/04
Foreign References:
| CN109685116A | 2019-04-26 | |||
| CN110287357A | 2019-09-27 | |||
| CN111105013A | 2020-05-05 |
Other References:
YI ZILI; ZHANG HAO; TAN PING; GONG MINGLUN: "DualGAN: Unsupervised Dual Learning for Image-to-Image Translation", 2017 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV), IEEE, 22 October 2017 (2017-10-22), pages 2868 - 2876, XP033283153, DOI: 10.1109/ICCV.2017.310
Attorney, Agent or Firm:
BEIJING ZHONG XUN TONG DA INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: