Image Classification Method Based on Improved KNN Algorithm

As the development of machine vision technology, artificial intelligence algorithms are gradually popularized for identifying images. However, traditional KNN algorithm actually costs too much time when classifying images, which is not qualified to actual application scenes. An improved algorithm is...

Full description

Saved in:
Bibliographic Details
Published inJournal of physics. Conference series Vol. 1930; no. 1; pp. 12009 - 12013
Main Authors Fan, Zhu, Xie, Jia-kun, Wang, Zhong-yu, Liu, Pei-Chen, Qu, Shu-jun, Huo, Lei
Format Journal Article
LanguageEnglish
Published Bristol IOP Publishing 01.05.2021
Subjects
Algorithms
Artificial intelligence
Image classification
Machine vision
Physics
Testing time
Online AccessGet full text
ISSN1742-6588
1742-6596
1742-6596
DOI10.1088/1742-6596/1930/1/012009

Cover

More Information
Summary:As the development of machine vision technology, artificial intelligence algorithms are gradually popularized for identifying images. However, traditional KNN algorithm actually costs too much time when classifying images, which is not qualified to actual application scenes. An improved algorithm is proposed in the paper. The test time has been greatly shortened and the efficiency of KNN algorithm is improved by increasing the screening of data sets. By setting STM32F103 as master control and OV7670 as camera, actual detection of volleyball, football, and basketball was carried out after test environment was set up. And the test time is shorter compared with that of general KNN algorithm. At the same time, the identification accuracy is high, which indicates that the method has good practicability.
Bibliography:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISSN:1742-6588
1742-6596
1742-6596
DOI:10.1088/1742-6596/1930/1/012009