Paper Info
Title
Automatic Labeling of Sensor Data Based on Object Tracking and Recognition
Abstract
The number of single-person households of seniors is increasing nowadays in many developed countries. In such kind of super-aging societies, solitary death is becoming a social problem. To solve the problem, we may use a smart home system to monitor the resident with cameras or voice recorders. But such devices have a big privacy issue. As an alternative solution, we are trying to develop a not-so-smart system. Specifically, the system considered in this study is a sparse array of infrared sensors. The sensor output is binary, and its state depends heavily on many factors (e.g., location, activity, and orientation of the subject). Thus, to define the "labels" of the data is not easy. In this paper, we study sensor data labeling based on video data. The rationale is that video data can help us to define/modify the labels both in real-time and afterward. Also in our sensor experiments, for the purpose to examine the infrared sensor recognize the activity strength for hand clapping motion as the first step. That is the reason, also we set the task to recognize the speed of hitting the hand by using the video data. So that during experiment a subject can move around and take some specified actions without having to stop frequently to mark the data.
Year
DOI
Venue
2018
10.1109/SCIS-ISIS.2018.00083
2018 Joint 10th International Conference on Soft Computing and Intelligent Systems (SCIS) and 19th International Symposium on Advanced Intelligent Systems (ISIS)
Keywords
Field
DocType
Image processing, activity recognition, location estimation, smart home, machine learning
Computer vision,Computer science,Video tracking,Artificial intelligence,Machine learning
Conference
ISSN
ISBN
Citations 
2377-6870
978-1-5386-2634-4
0
PageRank 
References 
Authors
0.34
2
2
Name
Order
Citations
PageRank
Tatsuya Hanyu110.72
Qiangfu Zhao221462.36