Related Experiment Video
Updated: Jan 11, 2026
08:05
Design and Analysis for Fall Detection System Simplification
Published on: April 6, 2020
11.1K
MIEF-Net: multimodal image-enhanced fusion network for intelligent fall risk prediction
Qizheng Zhao1, Ruiyuan Wu1, Manting Chen1
1Intelligent Sensing and Proactive Health Research Center, School of Public Health (Shenzhen), Shenzhen Campus of Sun Yat-sen University, Shenzhen, 518000, China; Shenzhen Key Laboratory of Pathogenic Microbes and Biosafety, Shenzhen Campus of Sun Yat-sen University, Shenzhen, 518000, China.
Summary
This study introduces an advanced deep learning method using wearable sensors to predict fall risk in older adults. The novel approach significantly improves accuracy, aiding in preventive geriatric care.
Area of Science:
- Gerontology
- Biomedical Engineering
- Artificial Intelligence
Background:
- Falls in older adults pose significant health risks, impacting quality of life.
- Early fall risk prediction and intervention are crucial for mitigating negative health outcomes.
- Inertial Measurement Unit (IMU) devices offer a portable solution for automated fall risk assessment.
Purpose of the Study:
- To propose a novel image-enhanced dual-stream deep learning approach for fall risk prediction using IMU-based gait analysis.
- To innovatively fuse spatial-temporal gait representations derived from IMU signals.
- To enhance the accuracy and clinical practicality of fall risk assessment.
Main Methods:
- Transforming raw IMU signals into Gramian Angular Field (GAF), spectrogram, and Markov Transition Field (MTF) images.
- Integrating recurrent neural networks (RNN) for sequential data and convolutional neural networks (CNN) for spatial features.
- Employing a Transformer-based multi-head attention mechanism for adaptive fusion of multi-modal gait data.
Main Results:
- The proposed approach achieved high performance metrics: Accuracy (0.9712), Sensitivity (0.9697), and F1-score (0.9708).
- Demonstrated superior performance compared to existing state-of-the-art models in fall risk prediction.
- The dual-stream design effectively identified subtle gait abnormalities missed by conventional methods.
Conclusions:
- The image-enhanced dual-stream deep learning model significantly advances fall risk prediction accuracy using wearable IMU sensors.
- This multi-modal fusion approach offers a clinically practical solution for preventive geriatric care.
- The study highlights the potential of advanced AI techniques in enhancing the safety and well-being of older adults.
