Reasons for using phase space in ECG diagnosis
The ECG monitoring system being developed by AVISS has major features in its elaborately designed filter and phase space based on the Hilbert transform.
AVISS's filter technology can be found at the following link posted on LinkedIn.
Briefly introduce the electrocardiogram-related projects that AVISS is working on.
In Fig. 1, the data analysis server is well known as the backend.
The ecg data sent from the cloud to mobile data is an electrocardiogram diagnosis report.
Although everything shown in Fig. 1 can be integrated on a mobile device, the reason for using a cloud system is that it is device independent.
This is because it is difficult to maintain saved data when the device is changed.
Another reason to use Cloud is schematized in Fig. 2.
The cloud built by AVISS can be used by the general public or small hospitals.
This system can be operated directly in the cloud by national health insurance, mobile electrocardiogram device manufacturers, large hospitals, and telecommunication companies.
The electrocardiogram-related projects underway at AVISS can be broadly divided into four categories.
1.backend s/w
2. Mobile ECG acqusition and transfer to cloud
3. Clinical testing
4. Obtaining FDA approval
Of the four above, the backend s/w part has been completed, and the remaining three are matters that need to be progressed in the future.
Since the backend s/w uses the PTB database, clinical testing of ECG data using actual mobile devices is a difficult task for AVISS. This is because the help of experts who can determine errors in diagnostic results for various cardiopathies is needed.
Diagnosis using mobile devices such as smartwatches is currently accepted as a medical device in many countries, so diagnostic functions can be used with limitations.
In addition, there is a problem that it is difficult to receive help from ECG-related diagnostic experts, as many cardiologist are opposed to ECG diagnosis using smartwatches, recognizing it as a form of telemedicine.
Due to these problems, it is difficult for AVISS to proceed with the project independently.
A way to make the project more accessible is to collaborate with existing mobile ECG monitoring companies or smartwatch manufacturers, but it cannot be ignored that it is difficult for them to easily change the direction of their project.
This situation will result in a truly wasteful outcome of following the path taken by existing companies.
Some companies appear to have achieved commercialization with investments of around $50 million. In the case of Google, it invested $2 billion to acquire a company and begin development, but it appears that it has not yet achieved tangible results.
It is clear that monitoring ECG based on a smartwatch like this is a very difficult project.
Several individuals or companies looking to invest have asked me this question.
1.???? Almost all companies are developing using AI, so why do we need the algorithm you boast about?
The three elements of AI can be said to be computers, data, and algorithms.
If you give AI ECG data, what process will it learn?
Computer calculations ultimately draw conclusions based on the logic of 0 and 1, so AI does not seem to require a separate algorithm for completely logical problems.
Are ECG signals logical?
If we give AI a specific signal and ask it to analyze frequencies, can it logically resolve the exact result?
I am not an AI expert, so I don't know the exact procedure.
However, some process will be needed to train AI.
That's the algorithm. If you give AI the correct algorithm, it will produce appropriate results.
Therefore, the AVISS algorithm using phase space can be a very useful tool for AI.
2.???? What are the advantages of the algorithm you are boasting about?
AVISS has three patents using phase space.
ECG signals have no phase information. Phase information of the electrocardiogram signal can be easily obtained through the Hilbert transform.
You can configure the phase space using the Hilbert transform and accurately find the R peak using AVISS' special filter. Mobile ECG measuring devices such as smartwatches provide single-lead ECG signals, so accurately finding the R peak is of utmost importance.
In cardiovascular disease, measurement of QT prolongation from single-lead electrocardiogram signals is subject to error with the currently existing methods. However, AVISS's patent using phase space has very high accuracy both in theory and in practice.
In the case of arrhythmia, we have seen several startups attempt smartwatch-based arrhythmia diagnosis but eventually switch to multi-lead systems.
AVISS perfectly responds to arrhythmia by introducing a filter to detect atrial fibrillation.
Regarding AV blocks, data from actual patients has not been secured, so it is still unconfirmed, but we do not believe there will be any major difficulties in responding to this issue in the future.