MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641100374 A) filed by Vignans Foundation For Science, Technology And Research on August 19, 2026, for A Deep Learning-Based Predictive Model For Binary And Multiclass Heart Disease Classification In Iomt.
Inventors include Mr. Peddipaga Rahul; Mr. K. Raghavendra Murali; Mr. Irfan Sayyad; Mr. A. N. Harshith Vardhan; and Dr. Gayatri Ketepalli.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: ABSTRACT [0015] The present invention discloses a Deep Learning-Based Predictive Model for Binary and Multiclass Heart Disease Classification integrated with the Internet of Medical Things (IoMT) to enable accurate, real-time detection and monitoring of cardiovascular diseases. The proposed framework combines advanced deep learning techniques with IoMT-enabled wearable devices to support both binary classification (presence or absence of heart disease) and multiclass classification (severity levels of heart disease). The framework begins with comprehensive data preprocessing, including duplicate removal, outlier detection, feature normalization, missing value imputation, and data standardization to improve model reliability. A Multi-Layer Perceptron (MLP) deep learning architecture is employed for disease prediction, utilizing sigmoid activation for binary classification and softmax activation for multiclass classification. Real-time physiological data collected from IoMT-enabled wearable sensors, including ECG, heart rate, blood pressure, and oxygen saturation, are transmitted through lightweight communication protocols such as MQTT and Bluetooth Low Energy (BLE) to a secure cloud infrastructure for continuous monitoring and AI-based inference. Experimental evaluation demonstrates that the proposed binary classification model achieves an accuracy of 94.6%, precision of 93.2%, recall of 95.8%, and an F1-score of 94.5%, while the multiclass model achieves an accuracy of 91.3% with consistently high classification performance across multiple disease severity levels. The integrated IoMT framework further achieves an average end-to-end latency of 2.3 seconds, enabling timely clinical decision support and remote patient monitoring. The proposed invention provides a scalable, intelligent, and clinically relevant healthcare framework suitable for hospitals, remote healthcare services, telemedicine platforms, cardiac monitoring centers, and smart healthcare ecosystems, thereby improving early diagnosis, continuous monitoring, and proactive management of heart disease. [0016] A deep learning-based predictive model for binary and multiclass heart disease classification is disclosed, integrated with the Internet of Medical Things to achieve accurate and timely detection of cardiovascular conditions. The framework applies comprehensive data preprocessing followed by a multi-layer perceptron architecture that employs sigmoid activation for binary outcomes and softmax activation for severity-level classification. Physiological signals including electrocardiogram, heart rate, blood pressure, and oxygen saturation are acquired from wearable sensors and transmitted via lightweight protocols to a cloud infrastructure for continuous inference. Experimental results show binary classification accuracy of 94.6 percent with precision of 93.2 percent, recall of 95.8 percent, and F1-score of 94.5 percent, while multiclass classification reaches 91.3 percent accuracy. End-to-end latency averages 2.3 seconds, supporting near real-time alerts. The system offers a scalable platform suitable for hospitals, remote monitoring services, telemedicine platforms, and smart healthcare ecosystems, thereby facilitating early diagnosis and proactive management of heart disease.
Disclaimer: Curated by HT Syndication.
