Open AccessA review of action recognition based on Convolutional Neural Network
Author(s) -
Jiaxin Yang,
Fang Wang,
Jianxiu Yang
Publication year - 2021
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1827/1/012138
Subject(s) - convolutional neural network , action recognition , computer science , action (physics) , artificial intelligence , pattern recognition (psychology) , machine learning , physics , quantum mechanics , class (philosophy)
At present, the development of video action recognition is very rapid in many fields, such as video understanding, intelligent monitoring, and human-computer interaction. However, there are some challenges in the development of action recognition, and researchers have tried to put forward some explorations. Convolutional neural network (CNN) is applied to action recognition, which improves the performance of action recognition. It is divided into 3 methods in this paper. In addition, C3D, Two-stream and I3D, three classic CNN algorithms, are reproduced. And their recognition rates are 72%, 78.0% and 97.6% respectively on the UCF101 dataset.