MUMBAI, India, March 13 -- Intellectual Property India has published a patent application (202641025049 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on March 3, for 'a privacy-preserving medical imaging system for detection of parkinson's disease from brain magnetic resonance imaging (mri) data.'
Inventor(s) include Dr. Dhivyaa C R; Mr. Rohith Reddy Gundadi; Mr. Bongu Sai Sushanth; and Mr. Bhumiredd Y Nivedh Reddy.
The application for the patent was published on March 13, under issue no. 11/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a privacy-preserving medical imaging system for detecting Parkinson's disease from brain magnetic resonance imaging (MRI) data is disclosed. The system includes an image preprocessing module configured to standardize MRI scans through resizing, normalization, pixel scaling, and tensor conversion for improved numerical stability. A multi-feature deep fusion engine extracts complementary diagnostic features using multiple convolutional neural networks, including a residual connection network for hierarchical representation learning, an inception-based network for multi-scale structural analysis, and a shallow convolution network for localized texture and edge detection. Extracted feature maps are concatenated in a fusion module to generate a unified deep feature representation, which is processed by a classification network to produce a diagnostic output indicative of Parkinson's disease. The system further incorporates a federated aggregation-proximal training controller enabling distributed local training across hospital or medical client devices without transferring raw patient MRI data. Locally trained parameters are aggregated using weighted averaging with proximal regularization to reduce divergence under heterogeneous datasets, thereby improving convergence stability, diagnostic accuracy, and patient data privacy in distributed medical environments."
Disclaimer: Curated by HT Syndication.
