MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641050041 A) filed by Andhra University, Visakhapatnam, Andhra Pradesh, on April 20, for 'a dual-backbone swin transformer and efficientnet-based framework for accurate prediction of alzheimer's disease stages.'
Inventor(s) include M K V Anvesh; and Dr. Prajna Bodapati.
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 present invention discloses a system and method for accurate prediction of Alzheimer's disease stages using a dual-backbone deep learning framework integrating a Swin Transformer and an EfficientNet architecture. The system processes neuroimaging data, including magnetic resonance imaging, through a preprocessing module that performs normalization, segmentation, and data enhancement. The preprocessed data is then simultaneously analyzed by the Swin Transformer to capture global contextual features and by the EfficientNet model to extract fine-grained spatial features. An adaptive feature fusion mechanism combines the outputs of both backbones into a unified representation, which is subsequently processed by a classification module to predict stages of Alzheimer's disease, including normal cognition, mild cognitive impairment, early-stage Alzheimer's disease, and advanced Alzheimer's disease. The proposed framework improves diagnostic accuracy, enhances feature representation, and supports clinical decision-making through reliable and interpretable predictions, thereby enabling early detection and effective management of neurodegenerative conditions."
Disclaimer: Curated by HT Syndication.
