MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126590 A) filed by S. Hrushikesava Raju; and Koneru Lakshmaiah Education Foundation on October 23, 2026, for A Pulmonary Arterial Hypertension Detection System Using Multimodal Clinical Data And A Cross-Attention Transformer.
Inventors include N. Sravanthi; S. Adinarayana; Arunreddy Pothireddy; C. Lakshminatha Reddy; S. Hrushikesava Raju; K. Hari; Jyosyula Satya Venu Gopala Krishna; S. Sherisha; T. Praveen; and Lakshmi Ramani Burra.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Due to high arterial pressure, pulmonary artery hypertension is a cardiovascular disease that results in high morbidity and high mortality if it is not detected in a very early stage. The diagnostic models using multi-modal should ensure a noninvasive approach and early detection. The automatic functioning of PAH using a cross- attention transformer is recommended over multi-modal data such as ECG signals, CT images, and MRI. The cross-attention transformer processes heterogeneous data streams using dedicated encoders: a CNN for imaging and a transformer for ECG signals. This model helps to fuse features from multimodal. The model observes intrinsic relationships, such as artery dilation in CT, right ventricular hypertrophy in MRI, and Cardiac strain in ECG, which helps to facilitate early and accurate detection. Compared to specific traditional methods, individual CNN and ViT, the multi-model cross-attention transformer achieves significantly better performance, specificity, and sensitivity. The cross-attention transformer not only ensures feature alignment but also interpretability with attention maps. This model is recommended as a non-invasive, robust approach for early PAH detection.
Disclaimer: Curated by HT Syndication.
