MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202631086334 A) filed by Narula Institute Of Technology on July 14, 2026, for “robust Arrhythmia Classification: Ai-Powered Real-Time Ecg Signal Diagnostic System Using Signal Processing And Deep Learning”.
Inventors include Asutosh Singh; Himadri Das; Sudipa Kower; Avik Dhar Chowdhury; and Dr. Sangita Roy.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: ABSTRACT This patent discloses a hardware-integrated system and method for real-time electrocardiogram (ECG) conditioning and robust multi-class arrhythmia classification deployed on low-power edge platforms. An analog-to-digital converter (ADC) node continuously digitizes raw analog biopotentials into 1D digital arrays, which are immediately cleaned via a phase-preserving 0.5–45 Hz Butterworth bandpass filter and normalized using dynamic Z-score scaling. A real-time temporal engine executes a local mathematical maximum search to locate R-peaks, prompting a window tracker to slice standardized, fixed-length 200-point heartbeat vectors centered on each peak. A lightweight 6-layer 1D-Convolutional Neural Network (1D-CNN) processes these vectors directly along a single temporal axis, using an inverse class-frequency weight penalty matrix embedded within its loss function to correct clinical data imbalances and classify heartbeats into five standard AAMI EC57 categories. An automated logging module serializes the diagnostic outputs with real-time hardware timestamps into relational databases.
Disclaimer: Curated by HT Syndication.
