MUMBAI, India, Aug. 31 -- Intellectual Property India has published a patent application (202641100578 A) filed by B. Rajani; and Dr. D. Vivekananda Reddy on August 20, 2026, for Segmentation-Guided Vision Transformers With Quantum-Inspired Optimization For Lung Cancer Detection Using Ct Images.
Inventors include B. Rajani; and Dr. D. Vivekananda Reddy.
The application for the patent was published on August 28, 2026, under issue no. 35/2026.
Abstract: The current invention reveals an automated nodule classification computer-based system and a method of lung nodules based on the use of computer tomography (CT) images. The system includes a processor and non-transitory memory that is set up to run a U-Net- based segmentation network to produce a lung region mask to identify and separate lung tissue and surrounding anatomical structures. The image of the segmented lung is divided into non-overlapping blocks and passed through a Vision Transformer encoder using multi-head self-attention to obtain contextual feature representations. An optimization framework based on quantum inspiration is applied, whereby a Quantum Genetic Algorithm is used to carry out feature selection by quantum-bit representation, probabilistic state updates, and Quantum-inspired Particle Swarm Optimization is used to tune hyperparameters. The optimized features are availed to a classification module so as to identify benign or malignant lung nodules. The invention offers less redundancy of features, better convergence of global optimization, better computational performance and better diagnostic accuracy to real time computer-aided lung cancer detection.
Disclaimer: Curated by HT Syndication.
