Detecting Healt Events From Compressed Sensed Signal
The invention makes it possible to automatically detect significant characteristics in a compressed physiological signal (by means of the compressive sensing technique) without necessarily requiring decompression. Furthermore, the characteristics so determined can be used to assess the state of health of a subject.
The invention allows to analyze physiological signals, and not only, acquired with sensors that directly record a compressed version of the signal, without additional computing resources and with low power consumption, a fundamental characteristic for battery-powered mobile devices. It allows to obtain information related to signal anomalies and / or events related to the health of a subject with a lower cost compared to decompression and subsequent analysis of the signal. Once an anomalous event has been determined, it is possible to send / save data for further verification using traditional methods. It further reduces energy consumption and memory usage in the device by limiting the transmission / storage of data to those related to abnormal events. Furthermore, the compressed signal preserves all the information contained in the original signal, which is not always true with traditional methods.
- Wearable Devices for Healthcare;
- Acquisition of physiological signals directly in compressed form;
- Reduction of energy consumption;
- Reduction of memory requirements;
- Simple computation;
- Real-time implementation.