Home >

news Help

Publication Information

English:User Adaptation of Convolutional Neural Network for Human Activity Recognition 
Japanese: Shinya Matsui, 井上 中順, Yuko Akagi, Goshu Nagino, 篠田 浩一.  
English: Shinya Matsui, Nakamasa Inoue, Yuko Akagi, Goshu Nagino, Koichi Shinoda.  
Language English 
Journal/Book name
English:2017 25th European Signal Processing Conference (EUSIPCO) 
Volume, Number, Page         pp. 753-757
Published date Oct. 26, 2017 
Conference name
English:European Signal Processing Conference (EUSIPCO) 
Conference site
English:the kos Island 
Official URL https://www.eusipco2017.org/
DOI https://doi.org/10.23919/EUSIPCO.2017.8081308
Abstract Abstract—Recently, monitoring human activities using smartphone sensors, such as accelerometers, magnetometers, and gyroscopes, has been proved effective to improve productivity in daily work. Since human activities differ largely among individuals, it is important to adapt their model to each individual with a small amount of his/her data. In this paper, we propose a user adaptation method using Learning Hidden Unit Contributions (LHUC) for Convolutional Neural Networks (CNN). It inserts a special layer with a small number of free parameters between each of two CNN layers and estimates the free parameters using a small amount of data. We collected smartphone data of 43 hours from 9 users and utilized them to evaluate our method. It improved the recognition performance by 3.0% from a userindependent model on average. The largest improvement among users was 13.6%. Index Terms—Human activity recognition, User adaptation, Convolutional neural network, Learning hidden unit contributions

©2007 Tokyo Institute of Technology All rights reserved.