Paper Info
Title
Human Activity Recognition by Means of Online Semi-supervised Learning
Abstract
Human activity recognition (HAR) is a reasonably recent field of study for the computer science community. It aims at automatically analysing ongoing events and extract their context from the captured data. The detection of human activities, such as walking, running, falling, or even cycling, allows for several heterogeneous applications, from surveillance systems to patient monitoring systems. Despite being a particularly active field of study in the past years, HAR still leaves many strategies left to explore and key aspects left to address. There are two main approaches in terms of data extraction: Video and sensors. The sensor approach is, however, the most promising, due to its extreme portability and unobtrusiveness. Most sensor-based HAR systems are trained in a static dataset with Supervised Learning techniques, generating a classification model with a relatively low error rate. However, these systems commonly ignore one of HAR's challenges, the difference of input signals produced by different people when doing the same activities. Consequently, as a user's movements drift from the generic, the system error increases. The activity classification method should therefore be able to generate adapted results for each different user. This article exhibits and discloses an under explored approach to this problem: By means of Online Semi-supervised Learning, An incremental technique capable of adapting the classification model to the user of the application by continuously updating it as the data from the user's own specific input signals arrives. This is possible due to the nature of Semi-supervised learning, which trains on both labeled and unlabeled data, making it possible to keep learning even after reaching its final user, without the need of any manual input.
Year
DOI
Venue
2016
10.1109/MDM.2016.93
2016 17th IEEE International Conference on Mobile Data Management (MDM)
Keywords
Field
DocType
Human Activity Recognition,Machine Learning,Online Semi-Supervised Learning
Data mining,Online machine learning,Activity recognition,Semi-supervised learning,Active learning (machine learning),Computer science,Supervised learning,Unsupervised learning,Artificial intelligence,Software portability,Data extraction,Machine learning
Conference
Volume
ISBN
Citations 
2
978-1-5090-0884-1
0
PageRank 
References 
Authors
0.34
5
2
Name
Order
Citations
PageRank
Hugo Louro Cardoso100.34
João Mendes-Moreira231729.50