Technology
Precision Human Motion Recognition
GaitMetrics excels in precise human motion recognition through the innovative use of mmWave (millimeter Wave) sensing technology. By analyzing Doppler shifts and point cloud data, our machine learning and statistical feature-based algorithms can accurately identify various human activities and postures. These algorithms can discern dynamic activities and the relative still postures of multiple individuals, including walking, standing, and lying down, ensuring comprehensive monitoring.
Customizable Human or Event Anomaly Detection
Leveraging the outcomes of our core technologies, we have designed algorithms to detect various anomalies tailored to your specific needs. For instance, in elderly care facilities, our system can detect falling activities. In sleep monitoring, it can identify changes in sleep patterns, which often indicate chronic health issues. These algorithms are available through APIs, facilitating seamless integration with existing IT systems.
Deep Learning-Based Vital Sign Detection
Our AIoT (AI+IoT) solution is highly sensitive to small movements, making it ideal for detecting chest movements caused by respiration and slight surface vibrations due to one’s heartbeat. However, the vital sign data obtained can be quite noisy. To address this, we have developed a cutting-edge denoising estimation neural network based on a deep diffusion model. This network significantly enhances the precision of vital sign detection by reconstructing the pulse ECG signal from radio sensing data.
Efficient Chip Processing and Secure SaaS System
Our comprehensive AIoT system integrates human detection algorithms on the radio sensing data and anomaly detection algorithms in the cloud.