MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202611116347 A) filed by Chandigarh University on September 29, 2026, for Self-Adaptive Explainable Deep Learning System For Evolutionary Lung Cancer Image Analysis.
Inventors include Aman Kataria; and Sonam Khattar.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The present invention discloses a self-adaptive explainable deep learning system for lung cancer image analysis. Digitized lung tissue images are standardized and processed through parallel convolutional and transformer-based feature extraction modules to obtain spatial and contextual features. An adaptive attention-based fusion module dynamically weights the extracted features, and an evolutionary feature optimization module iteratively selects and refines diagnostically relevant feature configurations. A hierarchical inference engine performs benign–malignant classification, lung cancer subtype classification, and risk stratification. An explainable artificial intelligence layer generates tissue-region heatmaps, feature-attribution information, and confidence scores corresponding to the diagnostic result. Low- confidence predictions and clinician-validated outcomes are processed through a controlled continuous learning mechanism that updates feature weights, feature configurations, and decision thresholds. Version control, audit logging, validation safeguards, and rollback maintain traceability and controlled deployment of updated model states.
Disclaimer: Curated by HT Syndication.
