Intellectual Property India Publishes Patent Application for Unified Artificial Intelligence System For Simultaneous Multi-Disease Detection From Chest X-Ray Images Using Multi-Label Deep Learning On Edge Hardware

MUMBAI, India, June 26 -- Intellectual Property India has published a patent application (202621060784 A) filed by Parul University Parul Institute Of Engineering And Technology on May 13, 2026, for Unified Artificial Intelligence System For Simultaneous Multi-Disease Detection From Chest X-Ray Images Using Multi-Label Deep Learning On Edge Hardware.

Inventors include Bharat Tank; Mitul Patel; and Soumya Das.

The application for the patent was published on June 19, 2026, under issue no. 25/2026.

Abstract: UNIFIED ARTIFICIAL INTELLIGENCE SYSTEM FOR SIMULTANEOUS MULTI-DISEASE DETECTION FROM CHEST X-RAY IMAGES USING MULTI- LABEL DEEP LEARNING ON EDGE HARDWARE A computer-implemented artificial intelligence system executes entirely on an offline embedded GPU computing device to perform automated multi-disease diagnosis from chest radiograph images. The system comprises an image ingestion module accepting DICOM, JPEG, PNG, BMP, or TIFF formats; a preprocessing module applying contrast-limited adaptive histogram equalization, spatial normalization, and channel standardization; a multi-label deep learning inference engine employing a convolutional neural network with independent sigmoid output layer executing a single forward pass to simultaneously generate probability scores for twelve or more disease classes including COVID-19, bacterial pneumonia, viral pneumonia, tuberculosis, lung cancer, lung nodules, pulmonary edema, atelectasis, pleural effusion, pulmonary fibrosis, emphysema, and normal condition; a per-disease visual explainability module generating independent Gradient-weighted Class Activation Maps for each disease class; a severity scoring module assigning quantified severity ratings based on probability magnitude and spatial activation extent; a triage priority engine evaluating all disease outputs to assign holistic case-level priority; and an automated report generator producing structured clinical reports with per-disease findings, visual heatmaps, severity ratings, and triage priority within two seconds, all without transmitting patient data externally.

Disclaimer: Curated by HT Syndication.