MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641049622 A) filed by Btp Madhav; and Koneru Lakshmaiah Education Foundation, Vaddeswaram, Andhra Pradesh, on April 18, for 'an ensemble multi-scale model to detect the multiple brain diseases using mri/ct images.'
Inventor(s) include Madhavi Garimella; P. Vidya Sagar; and B T P Madhav.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The current invention pertains to a multi-scale intelligent system, that is an ensemble, in the detection of multiple brain diseases, based on the Magnetic Resonance Imaging (MRI) and Computed Tomography (CT) images. The invention will offer a reliable and automated computer-aided diagnostic framework that can detect the different neurological disorders at an early stage through the combination of a sophisticated image processing and deep learning methods. It is a proposed system that involves an image preprocessing module that improves the neuroimaging data with a normalization strategy, denoising strategy, and anatomy alignment strategy before a multi-scale feature extraction mechanism is implemented that aims to extract both the fine-grained local features and global structural patterns of brain tissues. This is achieved by using an ensemble learning architecture made up of various deep learning models such as convolutional and attention-based networks to enhance robustness and diagnostic accuracy in a variety of heterogeneous imaging datasets. The system also builds in prediction fusion plan that integrates the results of more than one model to come up with the correct multi-disease classification results. Surprisingly, a 2D-to-3D reconstruction module allows to analyze the structure of the brain in 3D and perform volumetric analysis of lesions to aid clinical interpretation, an explainability component marks the areas of disease impact. One of the neurological disorders that the invention can spot in a single framework includes brain tumor, and Alzheimer disease. Experimental analysis shows that there is better accuracy, lower false rates of detection and higher generalization when compared with the traditional single model methodologies. The proposed invention will offer a scalable and effective approach to the intelligent neuroimaging analysis that will help in supporting early diagnosis, clinical decision-making, and sophisticated healthcare applications."
Disclaimer: Curated by HT Syndication.
