Image classification method and device based on differential network structure search

The invention provides an image classification method and device based on differential network structure search. The method comprises the following steps: acquiring an image to be classified; inputting the to-be-classified image into an image classification model to obtain an image classification re...

Full description

Saved in:
Bibliographic Details
Main Authors LI XING, HUANG BAIQIAO, GU QIJIA, ZHANG PENG, WANG KUNFU, ZHANG PENGYI, SU JIAN, FENG WEI
Format Patent
LanguageChinese
English
Published 27.09.2022
Subjects
Online AccessGet full text

Cover

Loading…
More Information
Summary:The invention provides an image classification method and device based on differential network structure search. The method comprises the following steps: acquiring an image to be classified; inputting the to-be-classified image into an image classification model to obtain an image classification result; the image classification model carries out classification based on a differentiable network structure search mode, and an importance index when the differentiable network structure search mode determines the importance of candidate operations in the search process is jointly determined by the number of times of sampling training of the candidate operations and the accuracy of the candidate operations on a verification set. According to the method, a new index is provided to judge the importance of the candidate operation based on the statistical information of the candidate operation in the network structure search process, and the importance index used for determining the importance of the candidate operatio
Bibliography:Application Number: CN202210646499