MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641112325 A) filed by Pragati Engineering College on September 18, 2026, for Ai-Based Ecg Anomaly Detection Using Deep Learning.

Inventors include Mrs. Lakshmi Viveka K; Dr. Manjula Devarakonda Venkata; Mr. Srikanth Kosaraju; and Ms. Rubia Tasneem.

The application for the patent was published on September 25, 2026, under issue no. 39/2026.

Abstract: The present invention provides a system and method for automated detection and classification of cardiac anomalies based on electrocardiogram (ECG) signals, which are wearable and provide real-time detection and classification. The ECG signals obtained from one or more electrodes are pre-processed for noise and baseline wander removal, and divided into individual heartbeat cycles. This hybrid deep-learning framework uses a Convolutional Neural Network (CNN) and a Long Short-Term Memory (LSTM) network to learn complementary morphological and temporal information from the ECG signal. An integrated attention mechanism is used to assign the extracted features with relevance weights, indicating the regions in a waveform that most influence the classification decision, and an interpretable annotation of the segment of the ECG is created. The signal is classified as being from a normal or anomaly class and a confidence score is assigned. The trained CNN-LSTM-attention model is compressed through pruning and/or quantization and then directly executed on an embedded processor to provide real-time inference without the need of cloud processing on the device. The system will produce local and/or wireless alert when an anomaly is detected that is above a predetermined confidence level and will send the classification and annotated waveform to an external device that has been paired with it. The hybrid feature extraction, attention-based interpretability, model compression, and edge inference are integrated into a single system, resulting in continuous, computationally efficient and explainable ECG anomaly monitoring for wearable and remote cardiac monitoring applications.

Disclaimer: Curated by HT Syndication.