Zhaozhen Tong
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Zhaozhen Tong has a total of 7 co-authors. Here are the top co-authors by number of shared publications:
- Masahide Nakamura
Graduate School of Engineering, Kobe University, 1-1 Rokkodai-cho, Nada-ku, Kobe 657-8501, Japan
4 shared publications - Sinan Chen
Graduate School of Engineering, Kobe University, 1-1 Rokkodai-cho, Nada-ku, Kobe 657-8501, Japan
4 shared publications - Hsin-Yen Yen
School of Gerontology and Long-Term Care, Taipei Medical University, No. 250, Wuxing St., Taipei 11031, Taiwan
2 shared publications - Kumiko Ono
Graduate School of Medicine, Kobe University, 7-10-2 Tomogaoka, Suma-ku, Kobe 654-0142, Japan
2 shared publications - Shu-Chun Lee
School of Gerontology and Long-Term Care, Taipei Medical University, No. 250, Wuxing St., Taipei 11031, Taiwan
2 shared publications - Yuko Yamaguchi
Department of Nursing, Graduate School of Health Sciences, Kobe University, 7-10-2 Tomogaoka, Suma-ku, Kobe, Hyogo 654-0142, Japan
2 shared publications
Latest Publications
- Open Access
CP-BayesNE: Uncertainty-Aware Personalized Nutrient-Density Estimation from Food Images Using Conformal Prediction and Prior-Corrected Dietary Adaptation
Background/Objectives: Reliable longitudinal dietary assessment is a prerequisite for precision nutrition, important for studying nutrition-related changes in health and physical function, but conventional dietary recording can impose substantial user burden. AI-driven, food-image-based assessment can reduce this burden, although uncertainty in food recognition and inter-individual dietary differences may propagate to nutrient estimation. Methods:We developed CP-BayesNE, an uncertainty-aware framework that combines temperature-scaled food classification, conformal candidate sets, training-prior-corrected dietary-history personalization, and posterior-weighted nutrient estimation to support AI-driven personalized nutrition. Results: EfficientNetV2-S achieved 78.01% Top-1 and 93.67% Top-5 accuracy, while temperature scaling reduced expected calibration error from 0.11238 to 0.04345. At , regularized adaptive prediction sets reduced the average candidate-set size from 80.78 to 5.77 while maintaining 94.32% empirical category coverage. In FoodLogAthl-218, prior-corrected personalization reduced nutrient density mean absolute error (MAE) by 4.0–6.6% across energy and macronutrients over 51 users. A single-participant 92-meal personal dataset provided preliminary end-to-end feasibility evidence, while also revealing limitations, including slightly worse fat MAE and conservative nutrient intervals. Conclusions: These findings support uncertainty-aware personalized dietary assessment while emphasizing that category-level conformal validity does not directly imply accurate true dietary intake.Open AccessCP-BayesNE: Uncertainty-Aware Personalized Nutrient-Density Estimation from Food Images Using Conformal Prediction and Prior-Corrected Dietary Adaptation
Background/Objectives: Reliable longitudinal dietary assessment is a prerequisite for precision nutrition, important for studying nutrition-related changes in health and physical function, but conventional dietary recording can impose substantial user burden. AI-driven, food-image-based assessment can reduce this burden, although uncertainty in food recognition and inter-individual dietary differences may propagate to nutrient estimation. Methods:We developed CP-BayesNE, an uncertainty-aware framework that combines temperature-scaled food classification, conformal candidate sets, training-prior-corrected dietary-history personalization, and posterior-weighted nutrient estimation to support AI-driven personalized nutrition. Results: EfficientNetV2-S achieved 78.01% Top-1 and 93.67% Top-5 accuracy, while temperature scaling reduced expected calibration error from 0.11238 to 0.04345. At , regularized adaptive prediction sets reduced the average candidate-set size from 80.78 to 5.77 while maintaining 94.32% empirical category coverage. In FoodLogAthl-218, prior-corrected personalization reduced nutrient density mean absolute error (MAE) by 4.0–6.6% across energy and macronutrients over 51 users. A single-participant 92-meal personal dataset provided preliminary end-to-end feasibility evidence, while also revealing limitations, including slightly worse fat MAE and conservative nutrient intervals. Conclusions: These findings support uncertainty-aware personalized dietary assessment while emphasizing that category-level conformal validity does not directly imply accurate true dietary intake. - Open Access
Event-Driven Multimodal Sensing and Computing for Context-Aware Home Monitoring Using Stereo Vision and Dietary Event Anchoring
Real-world home monitoring requires sensing systems that can capture daily behaviour without continuous raw-video retention or excessive user burden. However, domestic environments present irregular activity timing, fragmented human presence, asynchronous multimodal events, and privacy-sensitive data management. This study proposes an event-driven multimodal sensing and computing framework for context-aware home monitoring using stereo vision and dietary event anchoring. The framework integrates stereo RGB-based three-dimensional human motion sensing, dining-zone-triggered meal image acquisition, runtime event orchestration, timestamp-based cross-modal synchronization, privacy-aware local storage, and large-language-model-assisted dietary context interpretation. Instead of continuously recording all sensor streams, the system activates and organizes sensing through human presence detection, debounce logic, cooldown-based session control, and dining-zone occupancy events. Meal-related events are used as contextual anchors to associate motion sessions and dietary observations into synchronized behavioural episodes. The prototype was deployed for 11 consecutive days in a real kitchen–dining environment, with the stabilized real-time monitoring phase evaluated from 11 to 14 February 2026. During this phase, the system generated 26 event-driven motion sessions and 51,165 captured pose frames, of which 25,925 were valid. Sustained active sessions accounted for 30.8% of all sessions but contributed 81.5% of captured pose frames, indicating that event-driven orchestration concentrated motion data within behaviourally meaningful activity windows. Eight meal-related records were obtained, seven of which overlapped with motion sessions, resulting in 87.5% meal-event overlap coverage. Structured pose outputs required approximately 550 kB/min, corresponding to about 33 MB/h of recorded pose data. LLM-assisted meal-image interpretation achieved a mean absolute percentage error of 25.44%, supporting its use for coarse dietary-context description rather than precise nutritional quantification. However, this result is interpreted only as evidence for coarse dietary-context description and not as validation of a precise nutritional or clinical dietary assessment method. These results demonstrate the system-level feasibility of transforming irregular domestic observations into structured, temporally indexed, and privacy-aware multimodal behavioural records for future home monitoring applications.
Open AccessEvent-Driven Multimodal Sensing and Computing for Context-Aware Home Monitoring Using Stereo Vision and Dietary Event Anchoring
Real-world home monitoring requires sensing systems that can capture daily behaviour without continuous raw-video retention or excessive user burden. However, domestic environments present irregular activity timing, fragmented human presence, asynchronous multimodal events, and privacy-sensitive data management. This study proposes an event-driven multimodal sensing and computing framework for context-aware home monitoring using stereo vision and dietary event anchoring. The framework integrates stereo RGB-based three-dimensional human motion sensing, dining-zone-triggered meal image acquisition, runtime event orchestration, timestamp-based cross-modal synchronization, privacy-aware local storage, and large-language-model-assisted dietary context interpretation. Instead of continuously recording all sensor streams, the system activates and organizes sensing through human presence detection, debounce logic, cooldown-based session control, and dining-zone occupancy events. Meal-related events are used as contextual anchors to associate motion sessions and dietary observations into synchronized behavioural episodes. The prototype was deployed for 11 consecutive days in a real kitchen–dining environment, with the stabilized real-time monitoring phase evaluated from 11 to 14 February 2026. During this phase, the system generated 26 event-driven motion sessions and 51,165 captured pose frames, of which 25,925 were valid. Sustained active sessions accounted for 30.8% of all sessions but contributed 81.5% of captured pose frames, indicating that event-driven orchestration concentrated motion data within behaviourally meaningful activity windows. Eight meal-related records were obtained, seven of which overlapped with motion sessions, resulting in 87.5% meal-event overlap coverage. Structured pose outputs required approximately 550 kB/min, corresponding to about 33 MB/h of recorded pose data. LLM-assisted meal-image interpretation achieved a mean absolute percentage error of 25.44%, supporting its use for coarse dietary-context description rather than precise nutritional quantification. However, this result is interpreted only as evidence for coarse dietary-context description and not as validation of a precise nutritional or clinical dietary assessment method. These results demonstrate the system-level feasibility of transforming irregular domestic observations into structured, temporally indexed, and privacy-aware multimodal behavioural records for future home monitoring applications.
- Open Access
Towards Sensor-Based Mobility Assessment for Older Adults: A Multimodal Framework Integrating PoseNet Gait Dynamics and InBody Composition
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 (, ), SMM (, ), and intracellular water (, ), highlighting the role of muscle strength in movement adaptability. Conversely, negative correlations were found with PBF () and VFA (, ), 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.Open AccessTowards Sensor-Based Mobility Assessment for Older Adults: A Multimodal Framework Integrating PoseNet Gait Dynamics and InBody Composition
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 (, ), SMM (, ), and intracellular water (, ), highlighting the role of muscle strength in movement adaptability. Conversely, negative correlations were found with PBF () and VFA (, ), 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. - Open Access
- 12 June 2025
- Published by MDPI AG in Healthcare
- Vol. 13 (12) , 1405
Vision-Based Assessment of Skeletal Muscle Decline: Correlating Gait Variance with SPPB Performance
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.
- 12 June 2025
- Published by MDPI AG in Healthcare
- Vol. 13 (12) , 1405
Open AccessVision-Based Assessment of Skeletal Muscle Decline: Correlating Gait Variance with SPPB Performance
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.