Recently, smart meters have been widely used as social infrastructure, making it easy to obtain electricity power consumption data. Electricity power consumption data has high potential in healthcare since it reflects various activities in daily life. We have been focusing on the potential of smart meters and studying frailty detection model using electricity power consumption data; frailty is a condition of reduced function and health in the elderly, also said to be reversible, which means the elderly can return to robust conditions by implementing appropriate interventions. Therefore, detecting frailty at an early stage and implementing intervention makes it possible to extend healthy life expectancy. Our previous research showed that frailty detection is possible with acceptable accuracy only using electricity power consumption data from smart meters. However, for real-world implementation, it is crucial to guarantee the robustness of the model performance. Moreover, implementing interventions after detecting frailty and ongoing monitoring is essential to extending healthy life expectancy. Given these points, we aim to guarantee the robustness of frailty detection model and verify the effectiveness of frailty detection model as a post-intervention monitoring tool. As a result of case study conducted in three regions, we demonstrated that the proposed frailty detection model has seasonal and regional robustness, showing acceptable performance in various seasons and regions. We also confirmed that it is possible to observe the elderly’s condition by tracking the outputs of the proposed model. Through our research, we showed the feasibility of frailty detection using smart meters and the possibility of a healthcare platform using smart meters.

With the widespread implementation of smart meter infrastructure, various applications have been proposed and implemented beyond energy usage monitoring and savings. In particular, there has been a growing interest in using smart meter data for daily life monitoring and remote care systems by visualizing human activities based on electricity, gas, and water usage. Previous research has demonstrated the feasibility of activity detection using an individual smart meter dataset. However, research on the effectiveness of using multiple smart meter datasets is limited. This paper proposes a model for Activities of Daily Living (ADLs) detection using multiple smart meter datasets (i.e., gas and electricity). To evaluate the performance of the proposed approach, a case study was conducted with 47 households for one month, collecting electricity data, gas data, and ground truth data about ADLs. The case study results showed that our proposed approach, which uses multiple smart meter datasets, enables more accurate activity detection.

Emotion detection from face is an important problem and has received attention from industry and academia. Although emotion recognition from videos has a very high performance, emotion recognition from a single image stays a challenging task. In this paper, we try to use information from videos to do emotion recognition on a single image. More specifically, we leverage contrastive loss for pretraining the network on the videos and experiment with different sampling methods to select consistently hard triplets for continual learning of the network. Once the embeddings have been trained, we test them on a standard emotion classification task. Our method significantly improves the performance of the models and shows the efficacy of self-supervision in emotion recognition.

With the increase of the number of the elderly, healthcare systems to support the daily life and wellbeing of the elderly attracted attention. Especially, frailty syndrome is one of the most significant challenges faced by many countries because of its high association with mortality and hospitalization. Recently, with the progress of ICT (Information and Communication Technology), many frailty detection models which use sensors were proposed. However, many of them require very high costs caused by the installation and management of sensors. Therefore, the objective of this study is to propose a machine learning-based frailty detection model using only electric power consumption data from smart meter, which uses no other devices such as sensors. Also, we examined the feasibility of our model through a case study, in which we have conducted on 24 elderly people. As a result of a cast study, we could detect frailty with 82% accuracy, 77% precision, 84% recall, and 80% f-score for a 2-class classification problem (frailty or non-frailty). The results of our study show that more elderly people can receive frailty diagnoses through smart meters. Moreover, since frailty is a reversible condition that could be restored to a healthy status with early and appropriate intervention, our model has potential to extend the healthy expectancy of the elderly.

This letter proposes a new method of estimating the location of a predominant source in an amplitude-panned stereo signal with two sources. When the conventional method of location estimation is applied to an amplitude-panned multi-source sound, a serious estimation error occurs due to interference between sources. To solve this problem, the proposed method includes an embedded source separation based on non-negative matrix factorization, which first determines the initial estimate of source location, and then computes the basis matrix of the source using the initial location estimate. In this way, the proposed method can perform the source separation without a training stage. The comparative evaluation confirms that the proposed method provides higher performance in location estimation than the conventional method with a training stage.

In this paper, an adaptive compressed sensing is proposed in order to enhance the performance of fast tetrahertz reflection tomography. The proposed method first acquires data at random measurement points in the spatial domain, and estimates the regions in each tomographic image where much degradation is expected. Then, it allocates additional measurement points to those regions, so that more data are acquired adaptively at the regions prone to degradation, thereby improving the quality of the reconstructed tomographic images. The proposed method was applied to the T-ray reflection tomography system, and the image quality enhancement by the proposed method, compared to the conventional method, was verified for the same number of measurement points.