FACE IMAGE CLUSTERING METHOD, APPARATUS AND DEVICE, AND COMPUTER-READABLE STORAGE MEDIUM

WIPO Patent Application WO/2023/283805

A1

The present application discloses a face image clustering method, apparatus and device, and a computer-readable storage medium, the method comprising: acquiring a face training image from a face database, and updating in sequence all layers of a basis matrix, deep coefficient matrix and activation function parameters of a fully connected neural network on the basis of the face training image; after the basis matrix, the deep coefficient matrix and the activation function parameters are updated, obtaining a BP-DGNMF model; acquiring a face image to be clustered, and inputting the face image to be clustered into the BP-DGNMF model, so as to determine a deep feature set corresponding to the face image to be clustered; and on the basis of the deep feature set, performing a clustering operation on the face image to be clustered so as to obtain a clustering result corresponding to the face image to be clustered.

CHEN WENSHENG (CN)
ZHAN ZIHAO (CN)

PCT/CN2021/106073

January 19, 2023

July 13, 2021

Click for automatic bibliography generation  

UNIV SHENZHEN (CN)

G06K9/62

CN110942108A	2020-03-31
CN109711358A	2019-05-03
CN107657204A	2018-02-02
CN113052239A	2021-06-29
US20120041906A1	2012-02-16

"Master Thesis", 1 December 2016, NATIONAL UNIVERSITY OF DEFENSE TECHNOLOGY, CN, article ZIHENG GAO: "Research on Large-scale Non-negative Matrix Factorization with Graph Regularization", pages: 1 - 68, XP093023785

CENFO INTELLECTUAL PROPERTY AGENCY (CN)

View/Download PDF      PDF Help

Previous Patent: WIRELESS COMMUNICATION METHOD AND APPARATUS, AND TERMINAL DEVICE

Next Patent: A MOISTURE RESISTANT TWO-COMPONENT ADHESIVE COMPOSITION