Title:
QUANTIZATION METHOD AND APPARATUS
Document Type and Number:
WIPO Patent Application WO/2024/032775
Kind Code:
A1
Abstract:
The present application provides a quantization method, applicable to an application scenario where two devices share a same quantization dictionary. According to the method, information to be quantized is divided into a plurality of quantization parts, and quantization solutions for respective quantization parts are provided. A representation format of the quantization dictionary is provided, one quantization dictionary may be represented by dictionary usage and dictionary content, and this representation format actually establishes a corresponding relationship between R quantization parts of the information to be quantized and M quantization sub-dictionaries. If a first sub-dictionary in the M sub-dictionaries corresponds to a first quantization part in the R quantization parts, the first sub-dictionary is used for quantizing the first quantization part or dequantizing information obtained by quantizing the first quantization part, and each dictionary element in the first sub-dictionary is a candidate of quantized information of the first quantization part. By adopting the representation format of the quantization dictionary, the overhead of quantization dictionary storage of a device can be reduced.
Inventors:
CHEN JIAXUAN (CN)
HANG HAICUN (CN)
HANG HAICUN (CN)
Application Number:
PCT/CN2023/112591
Publication Date:
February 15, 2024
Filing Date:
August 11, 2023
Export Citation:
Assignee:
HUAWEI TECH CO LTD (CN)
International Classes:
H04B7/06
Foreign References:
| CN109075851A | 2018-12-21 | |||
| CN107733569A | 2018-02-23 | |||
| CN107170020A | 2017-09-15 | |||
| US20190312664A1 | 2019-10-10 |
Other References:
张婷 (ZHANG, TING): "基于量化的近似最近邻搜索技术研究 (Non-Official Translation: Research on Quantization-Based Approximate Nearest Neighbor Search Technique)", 中国科学技术大学硕士学位论文 (MASTER’S THESES OF OF UNIVERSITY OF SCIENCE AND TECHNOLOGY OF CHINA), no. 09, 15 September 2017 (2017-09-15), pages 4 - 21
MAN CHEN, ZHONG YONG, LI ZHEN-DONG: "Multi--focus image fusion based on latent low• •rank representation combining low• •rank representation", JOURNAL OF JILIN UNIVERSITY(ENGINEERING AND TECHNOLOGY EDITION)., vol. 50, no. 1, 1 January 2020 (2020-01-01), XP093137905, DOI: ：10.13229/j.cnki.jdxbgxb20180805
MAN CHEN, ZHONG YONG, LI ZHEN-DONG: "Multi--focus image fusion based on latent low• •rank representation combining low• •rank representation", JOURNAL OF JILIN UNIVERSITY(ENGINEERING AND TECHNOLOGY EDITION)., vol. 50, no. 1, 1 January 2020 (2020-01-01), XP093137905, DOI: ：10.13229/j.cnki.jdxbgxb20180805
Attorney, Agent or Firm:
LONGSUN LEAD IP LTD. (CN)
Download PDF:
Previous Patent: COMPOUND AS VOLTAGE-GATED SODIUM CHANNEL INHIBITOR
Next Patent: COMPOUNDS AS BCL-2 INHIBITORS
Next Patent: COMPOUNDS AS BCL-2 INHIBITORS