RAM: Real Time Activity Monitoring with feature extractive training

作者：

Highlights：

• SVM inspired method for real time human activity monitoring.

• Generate non-predefined features.

• Classifier carries out feature extraction as well.

• Multiclass classification as one-against-pseudo class, no hyperplane generation.

• Robust in inter-activity detection consistency, for growing set of activity classes.

摘要

•SVM inspired method for real time human activity monitoring.•Generate non-predefined features.•Classifier carries out feature extraction as well.•Multiclass classification as one-against-pseudo class, no hyperplane generation.•Robust in inter-activity detection consistency, for growing set of activity classes.

论文关键词：Activity recognition,Real time,Support Vector Machines,Accelerometer,Assisted living

论文评审过程：Available online 20 June 2015, Version of Record 10 July 2015.

论文官网地址：https://doi.org/10.1016/j.eswa.2015.06.017