Anomaly Detection Based on Denoising Autoencoder and Variational Autoencoder with Obrid-Sensor

Abstract

Aging society is a concern worldwide. The number of older individuals living alone is increasing. Thus, if they face an accident or get ill at home, there is a risk that they may die from the lack of early detection. Hence, existing systems employ cameras to detect such dangerous situations. However, this system is difficult to use in places where privacy needs to be protected, like in a bathroom. Thus, we developed a system that can detect abnormal states using an Obrid-Sensor. This sensor can observe the individual's state without invading their privacy because the brightness information to be acquired is one dimension. A previous study proposed a fall detection method with an autoencoder by determining the individual's state from standing data. Various noises cause a loss of performance in the autoencoder. Therefore, use of a denoising autoencoder and variational autoencoder was proposed, which are more robust to noise and more expressive, respectively. Consequently, recall of all methods was 100 %, and fall detection could be performed without false negative. The denoising autoencoder was the most suitable for anomaly detection.

Fall Detection Using Time-Series Data with Obrid-Sensor

“…Thus, this system can be used for privacy-conscious monitoring. In our previous studies, we proposed fall detection with using Deep Neural Network (DNN), Support Vector Machine (SVM), and AutoEncoder (AE) methods (6)(7)(8) . These are methods to analyze single data when the subject is standing or falling.…”