Benefit that SQ-Library
Many of life-threatening symptom
start outside of hospital.
Finding such signs at early stage is crucial
for better treatment.
NIHA（Neural Intelligent Heart Analyzer）- Software library finding heart abnormalities that uses ECG
NIHA-HF (Neural Intelligent Heart Analyzer) is an AI-based stand-alone software that uses lead I ECG datasets to provide a categorical assessment of a possibility of heart failure.
By studying over 20,000 lead I ECG recordings from about 10,000 patients, NIHA-HF is able to distinguish HF patients with different NYHA stages from healthy controls, with over 90% accuracy. This software can be used for early detection of HF and monitoring HF exacerbation. Combining the use of advanced signal processing with deep learning solution, this software could be implemented not only in standard 12-lead ECG devices, but also in lead I (or similar to lead I) wearable ECG products.
NIHA-AF (NIHA Atrial Fibrillation) is an AI-based stand-alone software package, aimed at detecting AF in lead I of an ECG.
The software combines the use of time-frequency analysis and a neural network to determine whether a 10-second ECG segment contains an AF event. By having studied over 60,000 ECG recordings, it is now able to detect the presence of AF events with 98.7% sensitivity. This software can be used for AF event detection in pre-recorded ECGs as well as for daily AF monitoring.
Stress Index is an objective indicator to estimate and quantify the stress level using ECG or pulse signal.
Stress Index (SI) is introduced for quantifying the stress level and providing a feasible solution for stress level comparisons among individuals. Stress, resulting from physical or psychosocial stressors, can influence the autonomous nervous system (ANS), where both the sympathetic and parasympathetic nervous systems are involved in the regulation of functions. Heart rate variability (HRV) is a reliable mean to observe ANS, which can be an indirect biomarker for stress estimation. SI is computed using reciprocal of selected HRV indexes including time, frequency domain and nonlinear features of HRV. It enables daily monitoring of stress level for the prevention of lifestyle-related chronic diseases over long term.
ECG Authentication is an AI-based software using single-lead ECG as a biometric for personal authentication.
ECG is highly personalized and appealing for human identification, as they are non-intrusive, continuously available, can only be measured in live subjects, and can not be simply intercepted, or duplicated. This software enables fast authentication by using only 3 beats within 3 seconds, meanwhile achieves promising performance in recognition. The software has been tested with a single-lead ECG device and achieved average FAR and FRR below 10% on a group of 50 subjects. This performance is more promising on small groups of cohort (less than 30), having obtained both FAR and FRR under 5.00%.