Activity Recognition Based on DCNN and Kinect RGB Images

Tan Hsu Tan, Munkhjargal Gochoo, Hong Syuan Chen, Shing Hong Liu, Yung Fa Huang

Research output: Chapter in Book/Report/Conference proceedingConference contribution

1 Citation (Scopus)

Abstract

Recognizing daily activities of elderly people can detect abnormal activity of elder people living alone. Therefore, activity recognition has received much attention in recent years. This study aims to recognize daily activities of people by employing RGB activity images and deep convolutional neural network (DCNN). In this study, the Cornell Activity Dataset (CAD-60) collected by RGB-D camera via the Microsoft Kinect is employed to train the DCNN model. Then, the DCNN model is used to classify activity images. Experimental results of 4-fold cross validation show that the precision, recall, specificity, accuracy, and F1-score of 95.5%, 95.6%, 99.8%, and 99.6%, and 95.3%, respectively, are achieved. The result is superior to other existing systems, indicating the application potential of our work.

Original languageEnglish
Title of host publication2020 International Conference on Fuzzy Theory and Its Applications, iFUZZY 2020
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781728198125
DOIs
Publication statusPublished - Nov 4 2020
Event2020 International Conference on Fuzzy Theory and Its Applications, iFUZZY 2020 - Hsinchu, Taiwan, Province of China
Duration: Nov 4 2020Nov 7 2020

Publication series

Name2020 International Conference on Fuzzy Theory and Its Applications, iFUZZY 2020

Conference

Conference2020 International Conference on Fuzzy Theory and Its Applications, iFUZZY 2020
Country/TerritoryTaiwan, Province of China
CityHsinchu
Period11/4/2011/7/20

Keywords

  • Activity Recognition
  • CAD-60
  • Deep Convolutional Neural Network (DCNN)
  • RGB activity image

ASJC Scopus subject areas

  • Applied Mathematics
  • Control and Optimization
  • Logic

Fingerprint

Dive into the research topics of 'Activity Recognition Based on DCNN and Kinect RGB Images'. Together they form a unique fingerprint.

Cite this