Apparatus and method for validating labeled data generated by a prediction generator
Inventors
Jain, Rohit • Ramani, Anand • Murty, Kappagantula Gopalakrishna • H J, Yogisha • Nadapurohit, Sanjeev Shrinivas • Bharadwaj, Karthik K. • Ptaszek, Leon
Assignees
Interested in licensing this patent?
MTEC can help explore whether this patent might be available for licensing for your application.
Publication Number
US-12453518-B2
Publication Date
2025-10-28
Expiration Date
Abstract
An apparatus and method for validating, using user input, labeled data generated by a prediction generator. The apparatus includes at least a processor and a memory communicatively connected to the at least a processor. The processor receives at least a potential signal, generates, processed data using the at least a potential signal, trains a prediction generator on a plurality of labeled training data, wherein the plurality of labeled training data comprises the processed data associated with at least an annotation, generates, using the prediction generator, a plurality of labeled data as a function of dynamic data, displays the plurality of labeled data to a user interface of a graphical user interface, receives a user input associated with the plurality of labeled data, and retrains the prediction generator using the user input.
Core Innovation
The disclosed system and method validate labeled data generated by a prediction generator using user input. An intracardiac catheter and localization system provide potential signals and position signals to produce processed data. A prediction generator produces labeled data from the processed data for display through a graphical user interface. The user supplies confirmation or correction data, which is subsequently used to retrain the prediction generator.
The validation of labeled data is performed by evaluating catheter-derived signals against annotations to facilitate the iterative retraining of the prediction generator. The prediction generator utilizes a semi-supervised machine-learning model that includes a neural network. The system integrates signal processing, model-generated labeling, user validation, and database storage to refine the intracardiac labeled data.
Claims Coverage
The claims include four inventive features across two independent claims: an apparatus claim and a method claim directed to validating labeled intracardiac data.
Cardiac-signal acquisition and localization
An intracardiac catheter with a transducer detects a cardiac phenomenon and outputs a potential signal, and a localization system detects a position signal as a function of catheter location.
Generation and display of labeled data
A computing device receives the potential signal, generates processed data, trains a prediction generator on labeled training data comprising archive data or processed data associated with an annotation, and displays generated labeled data via a downstream graphical user interface.
Display and user validation of labeled data
A graphical user interface receives user input associated with the labeled data, wherein the input comprises a confirmation datum or a correction datum including a signal segment correction or a segment annotation correction.
Prediction generator retraining
The prediction generator is retrained using the received user input, wherein dependent claims further recite semi-supervised neural network models and storing corrections in a database.
The independent claims cover acquiring cardiac potential and position signals, generating and displaying labeled data, receiving user confirmation or corrections, and retraining the prediction generator using the input data. Dependent claims further recite model architectures, signal processing parameters, and correction storage.
Stated Advantages
Not explicitly described in patent.
Documented Applications
Validation of prediction-generator labels for intracardiac signals through a graphical user interface where user confirmation or correction of signal segments and segment annotations is used to retrain the prediction generator.
Interested in licensing this patent?