This study introduces a dual-proxy framework for household-adaptive, non-invasive shower detection using standard water-heater logs. The framework leverages proxy at two complementary levels: a feature-level proxy (washing_seconds) that captures washing duration, and a scheme-level proxy (proxy-driven training) that enables learning in periods without direct shower labels. The proxy feature (washing_seconds) serves as an indirect descriptor of washing behavior, enabling effective inference even under label scarcity. We investigated three research questions: (RQ1) the effectiveness of proxy features in improving shower detection, (RQ2) how proxy-driven evaluation identifies compact yet reliable feature subsets, and (RQ3) the robustness of these subsets in long-term, real-world scenarios. Experiments on two households showed that washing_seconds consistently improved discrimination (raising summer PR-AUC, lowering non-summer false alarms), and that compact subsets of only two or three features, anchored by the proxy feature, achieved stable performance across households. The evaluation represents an illustrative example based on two cooperating households, providing practical evidence of the framework's real-world applicability. Evaluation in real-world conditions confirmed robustness: representative subsets maintained micro PR-AUC 0.724-0.728, micro F1 0.66-0.69 (macro F1 0.55-0.58), and summer PR-AUC near 0.87, with generalization gaps within +/- 0.01 for discrimination and small positive shifts for F1 (+0.02-+0.05). These results demonstrate that proxy can function both as a feature and as a methodological principle, and that the proposed framework is model-agnostic and transferable to other learning architectures. It provides a foundation for adaptive, privacy-preserving smart home applications that can scale to broader household and healthcare contexts.

The acceleration of global population aging has driven a surge in demand for health monitoring among older adults. However, traditional mobility assessment methods mostly rely on invasive measurements or laboratory-grade equipment, making it difficult to achieve continuous monitoring in daily scenarios. This study investigated the correlation between dynamic gait characteristics and static body metrics to enhance the understanding of elderly mobility and overall health. A sensor-based framework was implemented, which utilizes the Short Physical Performance Battery (SPPB), combined with PoseNet (a vision-based sensor) for dynamic gait analysis, and the InBody bioelectrical impedance device for static body composition assessment. Key variables comprised the dynamic metric mean directional shift and static metrics, including skeletal muscle index (SMI), skeletal muscle mass (SMM), body fat percentage (PBF), visceral fat area (VFA), and intracellular water. Nineteen elderly participants aged 60–89 years underwent assessments; among them, 16 were males (84.21%), and 3 were females (15.79%), 50% were in the 80–89 age group, 95% did not live alone, and 90% were married. Dynamic gait data were analyzed for center displacement and horizontal directional shifts. A Pearson correlation analysis revealed that the mean directional shift positively correlated with SMI ((Formula presented.), (Formula presented.)), SMM ((Formula presented.), (Formula presented.)), and intracellular water ((Formula presented.), (Formula presented.)), highlighting the role of muscle strength in movement adaptability. Conversely, negative correlations were found with PBF ((Formula presented.)) and VFA ((Formula presented.), (Formula presented.)), suggesting that greater fat mass impedes dynamic mobility. This multimodal integration of dynamic movement patterns and static physiological metrics may enhance health monitoring comprehensiveness, particularly for early sarcopenia risk detection. The findings demonstrate the framework’s potential, indicating mean directional shift as a valuable dynamic health indicator. © 2025 by the authors.

Background: With the global population aging, the proportion of the elderly is increasing, leading to health challenges. The decline in the elderly's physical function raises their fall risk, which affects their health and burdens the healthcare system. Traditional fall risk assessment methods like Short Physical Performance Battery (SPPB) have limitations, while computer vision technology shows potential but also has drawbacks. Objective: This study aims to use computer vision technology to quantify the elderly's gait movement features, analyze their correlations with SPPB test scores and duration consumption, and explore a solution for long-term monitoring and more efficient fall risk assessment. Methods: Data from 19 elderly Japanese subjects, including SPPB test data and camera-captured body movement data, were analyzed. Python (Version 3.12.6) was used to obtain JSON data, calculate movement distances, and construct a comprehensive index. Correlation analysis and principal component analysis (PCA) were performed. Results: The variance mean indicator of the comprehensive index associated with movement distance had a significant negative correlation with the completion duration of Test 2 in SPPB, indicating that greater gait variability might be related to better physical vitality. PC1 (Muscle-Control Reserve) and PC2 (Learning-Fatigue Response) obtained from PCA had a positive relationship with the test duration. The comprehensive index had a positive but not highly significant correlation with test scores. Conclusions: This study analyzed the correlation between the elderly's gait movement features and SPPB test performance. It innovated in data collection and analysis methods. Future research can be improved by expanding the sample size, adding more parameters, and applying deep-learning techniques.

Understanding human intention is critical in healthcare, where direct communication might be limited. We developed a dynamic intention recognition system for elderly care, integrating visual data and conversational interactions. Building upon in-home fixed-point cameras and lightweight image recognition systems, facial, skeletal, and hand features were detected to monitor behavior changes. Image-based analysis alone is often insufficient to fully grasp the intentions behind observed actions. Therefore, we developed a multimodal system that combines (1) image-derived contextual cues and (2) dialogue with an interactive agent, analyzed through locally executable large language models (LLMs). Specifically, we utilized locally deployable models such as Meta-Llama and Llama 3.1 within a web-based interface built on AnythingLLM, enabling real-time conversational intention estimation without compromising data privacy. The developed system performs image-based context extraction, conversational prompting with embedded intentions, and time-windowed log analysis using local LLMs for chronological intention estimation and confirmation. Five participants were recruited to compare their self-reported intentions with those inferred by the system using document similarity metrics. The results demonstrated the feasibility of the system and its real-time, privacy-conscious performance in elderly home care. The system contributes to the advancement of human-agent collaboration and lays a foundation for adaptive, intention-aware support systems. © 2025 IEEE.

In recent years, percentage of the elderly living alone are also on the rise. Elderly living alone are likely to be socially isolated, and if the detection of illness or injury is delayed, it is likely to lead to dying alone. It is important to understand the daily health condition of such elderly people, but existing safety confirmation and elderly care services have the problem that elderly people have no choice but to report their health condition by themselves. In this study, we use an existing emotion recognition model to acquire and analyze emotion estimation results from images of facial expressions of people acting according to their health condition. Then, we conduct an experiment to see if machine learning can classify the video clips using the emotion estimation results as feature values and the health condition of the person as labeled data. Based on the experiments, it is possible to classify facial expressions by creating a model that is adapted to the individual. © 2025 IEEE.

In Japan, the prevalence of Mild Cognitive Impairment (MCI), a prodromal stage of dementia, has been increasing with population aging, becoming a significant social issue. MCI impairs memory, particularly episodic memory, imposing psychological and time-related burdens on both individuals and caregivers. Although several methods have been proposed to support memory by collecting and presenting daily life data, many rely on third-person perspectives or require complex device operations. Therefore, it is difficult for such individuals to operate existing systems independently. The goal of this paper is to support older adults with MCI in independently collecting and storing first-person perspective data, thereby facilitating self-initiated recall of episodic memories. Our key idea is to integrate a wearable device built with a Raspberry Pi and cloud storage. We propose a system that automatically captures first-person view images, stores them in the cloud, and enables search based on tags assigned through object recognition. This system aims to support self-initiated recall of episodic memories by allowing older adults to independently collect and retrieve contextual data from their own perspective. © 2025 IEEE.

Japan's aging population and declining birthrate have had a significant impact on emergency medical services(EMS). The rising number of emergency transports has led to longer response times. Some less urgent patient transfers additionally delay the provision of critical emergency care. The introduction of private EMS has been considered as a potential solution, but its effectiveness remains unverified. This study proposes a method to evaluate the impact of delegating less urgent patient transfers to private EMS on EMS quality, using data from Kobe City. Results show that over a year, the cumulative response time was reduced by approximately 170 million seconds, indicating the delegation of less urgent patient transfers to private EMS improves the quality of EMS. © 2025 IEEE.

In recent years, climate change, particularly global warming, has become a major concern, with global average temperatures continuing to rise. In response, the Japanese government has declared its commitment to achieving a zero-carbon society where the difference between CO2 emissions and absorption becomes zero. While energy conservation is crucial for achieving this goal, this research specifically focuses on energy-saving behaviors in general households. Previous research has proposed methods to promote energy-saving behaviors through dialogue between systems and users. However, these methods have several issues: they can only provide predetermined responses, they use costly 3D models, and they cannot determine whether users are at home. Considering these issues, this research proposes EcoWiz, a system utilizing large language models and Live2D. Additionally, the system implements presence detection and periodic notifications using motion sensors and Pub/Sub architecture. Based on the proposed method, we implemented the system and verified its operation through case studies. © 2025 IEEE.

In Japan, the number of elderly people and dementia patients is on the rise. They often face the issue of forgetting where they placed their belongings. When they are unable to remember the location of an item, not only do they themselves have to search for it, but their family members and caregivers are also involved in the search, increasing the burden on everyone involved. Therefore, reducing this burden has become a significant challenge. To address this issue, this study proposes a "Grasped Object Recognition and Recording Service"that utilizes cameras installed in the room and image recognition technology to identify and record objects grasped by individuals. The implementation focused on the recognition and recording of the grasped objects. As a result, the system successfully achieved a practical level of accuracy in recognizing and recording the objects held by people in the room. © 2025 IEEE.

Japan is experiencing rapid population aging, and the shortage of healthcare workers has become a serious issue. In parallel, the number of people with dementia is increasing, raising the importance of early detection and robust screening methods. To address these challenges, we developed the tablet-based drawing test app EVIDENT. In this study, we propose an early screening method for cognitive decline using drawing-process data from the Clock Drawing Test (CDT) and the Cube Copying Test (CCT) collected with these apps. Specifically, we build machine learning models that use process features as explanatory variables and scoring results as target variables. We also perform clustering of the drawing-process data to reveal drawing-process patterns in participants with cognitive decline and to summarize their characteristics. Through evaluation experiments, we demonstrated the effectiveness of the proposed models and extract process features that contribute to early screening. These results suggest that our approach can serve as a basis for supporting early detection in real-world operations, including group testing, while reducing burden at the point of care. In future work, we will expand the dataset, further improve screening accuracy, and pursue applications to diagnostic support integrated with EVIDENT. © 2025 IEEE.