MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126204 A) filed by Mr. Nagaraj V Telkar; Dr. Shivayogi Ullagaddi; Shreya M Chinnur; Sindhu Patil; Swati S Kushtagi; and Varsha A Mahendrakar on October 22, 2026, for Explainable Electrocardiographic Signal Analysis System Using Convolutional And Bidirectional Recurrent Neural Networks.
Inventors include Mr. Nagaraj V Telkar; Dr. Shivayogi Ullagaddi; Shreya M Chinnur; Sindhu Patil; Swati S Kushtagi; and Varsha A Mahendrakar.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The present invention an electrocardiographic signal analysis system comprises an input interface, a processor, a memory storing trained neural network parameters, and a display. The processor receives an electrocardiographic recording, reduces noise, normalizes signal amplitudes, and segments the recording into inputs compatible with a selected model. Stored models include a convolutional neural network and a convolutional network coupled to a bidirectional long short term memory network. The selected model generates scores for supported cardiac waveform categories. An explanation module applies gradient weighted class activation mapping or Shapley additive explanations to identify signal portions contributing to a selected output. The display presents the waveform, predicted category, associated score, and explanatory information. A symptom interface receives reported symptoms for presentation with the electrocardiographic result in a screening summary. An evaluation module compares model performance using classification metrics and noise robustness tests. The system provides explainable analysis of electrocardiographic recordings for screening support and subsequent professional review. (Accompanied Figure No. 1-2)
Disclaimer: Curated by HT Syndication.
