Methods and systems for determining position signals of electrodes using a retrained machine-learning model
A system for determining position signals of electrodes using a retrained machine-learning model includes at least a catheter including a plurality of electrodes configured to collect a plurality of potential signals and a magnetic sensor configured to collect magnetic data, and at least a computing device including a memory. The processor receives a first training set, wherein the first training set includes patient-agnostic data; receives a second training set, wherein the second training set includes patient-specific data, trains a mapping machine-learning model using the first training set, retrains the mapping machine-learning model using the second training set, receives at least a first signal, wherein the first signal includes a potential signal of the plurality of potential signal and the magnetic data, and generates, using the retrained machine-learning model, as a function of the at least a first signal, a first position signal for an electrode of the plurality of electrodes.
1. A system for determining position signals of electrodes using a retrained machine-learning model, the system comprising:
at least a catheter comprising:
a plurality of electrodes configured to collect a plurality of potential signals; and
a magnetic sensor configured to collect magnetic data; and
at least a computing device, wherein the computing device comprises:
a memory; and
at least a processor connected to the memory, wherein the memory contains instructions configuring the at least a processor to:
receive a first training set, wherein the first training set comprises patient-agnostic data;
receive a second training set, wherein the second training set comprises patient-specific data;
train a mapping machine-learning model using the first training set, wherein the mapping machine-learning model is configured to determine position signals for electrodes as a function of magnetic information and potential signals;
determine an accuracy score of the mapping machine-learning model based on a plurality of cohort feedback;
retrain the mapping machine-learning model based on the accuracy score using the second training set;
receive at least a first signal, wherein the first signal comprises a potential signal of the plurality of potential signals and the magnetic data; and
generate, using the retrained mapping machine-learning model, as a function of the at least a first signal, a first position signal for an electrode of the plurality of electrodes;
evaluate the retrained machine-learning model using point-wise system configured to analyze an accuracy of predicted positions or signals by evaluating an error between a true and predicted points in three-dimensional space.
2. The system of claim 1 , wherein:
the plurality of electrodes comprises:
a reference electrode located on a rigid portion of the catheter; and
a query electrode located on a flexible portion of the catheter; and
the first position signal is for the query electrode.
3. The system of claim 1 , wherein the mapping machine-learning model comprises a supervised deep learning model.
4. The system of claim 1 , wherein the memory contains instructions further configuring the processor to validate the first position signal by generating a cross-electrode model validation.
5. The system of claim 1 , wherein a location of the plurality of electrodes is computed as a function of at least the magnetic sensor and rigid inter-electrode distances.
6. The system of claim 1 , wherein retraining the mapping machine-learning model using the at least a second training set comprises retraining the mapping machine-learning model using single shot learning.
7. The system of claim 1 , wherein:
the plurality of electrodes comprises:
a reference electrode of first electrode type; and
a query electrode of second electrode type; and
receiving the at least a first signal comprises:
receiving a first potential signal from the reference electrode;
receiving a second potential signal from the query electrode; and
normalizing the second potential signal with respect to the first potential signal as a function of an electrode-type transformation model.
8. The system of claim 1 , wherein the at least a first training set comprises exemplary magnetic data, exemplary reference electrode potential signals, exemplary reference electrode position signals, and exemplary query electrode potential signals correlated to exemplary query electrode position signals.
9. The system of claim 4 , wherein the memory contains instructions further configuring the processor to transform potential signal readings across the plurality of electrodes using the cross-electrode model validation.
10. The system of claim 7 , wherein the electrode-type transformation model is a machine-learning model trained using training data correlating potential readings from the first electrode type to potential readings from the second electrode type.
11. A method for determining position signals of electrodes using a retrained machine-learning model, the method comprising:
receiving, by at least a processor, a first training set, wherein the first training set comprises patient-agnostic data;
receiving, by the at least a processor, a second training set, wherein the second training set comprises patient-specific data;
training, by the at least a processor, a mapping machine-learning model using the first training set, wherein the mapping machine-learning model is configured to determine position signals for electrodes as a function of magnetic information and potential signals;
determine an accuracy score of the mapping machine-learning model based on a plurality of cohort feedback;
retraining, by the at least a processor based on the accuracy score, the mapping machine-learning model using the at least a second training set;
receiving, by at least a catheter, at least a first signal, wherein the first signal comprises a potential signal of a plurality of potential signals and the magnetic data; and
generating, by the at least a processor, using the retrained machine-learning model, as a function of the at least a first signal, a first position signal for an electrode of the plurality of electrodes;
evaluating, by the at least a processor, the retrained machine-learning model using point-wise system configured to analyze an accuracy of predicted positions or signals by evaluating an error between a true and predicted points in three-dimensional space.
12. The method of claim 11 , wherein:
the plurality of electrodes comprises:
a reference electrode located on a rigid portion of the catheter; and
a query electrode located on a flexible portion of the catheter; and
the first position signal is for the query electrode.
13. The method of claim 11 , wherein the mapping machine-learning model comprises a supervised deep learning model.
14. The method of claim 11 , further comprising, validating, by the at least a processor, the first position signal by generating a cross-electrode model validation.
15. The method of claim 11 , wherein a location of the plurality of electrodes is computed as a function of at least a magnetic sensor and rigid inter-electrode distances.
16. The method of claim 11 , wherein retraining the mapping machine-learning model using the at least a second training set comprises retraining the mapping machine-learning model using single shot learning.
17. The method of claim 11 , wherein:
the plurality of electrodes comprises:
a reference electrode of first electrode type; and
a query electrode of second electrode type; and
receiving the at least a first signal comprises:
receiving a first potential signal from the reference electrode;
receiving a second potential signal from the query electrode; and
normalizing the second potential signal with respect to the first potential signal as a function of an electrode type transformation model.
18. The method of claim 11 , wherein the at least a first training set comprises exemplary magnetic data, exemplary reference electrode potential signals, exemplary reference electrode position signals, and exemplary query electrode potential signals.
19. The method of claim 14 , further comprising, transforming, by the at least a processor, potential signal readings across the plurality of electrodes using the cross-electrode model validation.
20. The method of claim 17 , wherein the electrode type transformation model is a machine-learning model trained using training data correlating potential readings from the first electrode type to potential readings from the second electrode type.