MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202541123486 A) filed by Saveetha Engineering College, Chennai, Tamil Nadu, on Dec. 8, 2025, for 'multimodal alzheimer's diagnostic system : cnn, grad-cam, and chatbot integration.'
Inventor(s) include Ms Archana. S. H; Dr. Surya. S; Mr. Sriram. S. S; Ms. Kaviyashree. S; and Mr. Augustine. J.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "This invention relates to a multimodal Alzheimer's diagnostic system that combines deep learning-based MRI analysis, visual explainability, and conversational interaction in a unified framework. The system utilizes a balanced dataset of approximately 12,000 brain MRI images distributed across four dementia stages-Non-Demented, Very Mild Dementia, Mild Dementia, and Moderate Dementia-and employs a transfer-learning ResNet18 architecture with a custom classifier head trained using cross-entropy loss with label smoothing, Adam W optimization, a ReduceLROnPiateau scheduler, and early stopping. During inference, an uploaded MRI scan is preprocessed and passed through the network to obtain a predicted dementia stage and class probabilities, while a Grad-CAM module generates a heatmap over the image to highlight regions most responsible for the prediction. A chatbot module then consumes the prediction and Grad-CAM context to provide natural-language explanations, answer Alzheimer's-related questions, and guide users through the implications of the result. By integrating classification, explainability, and interactive counseling, the invention offers an interpretable, patient-centric diagnostic tool that improves transparency and usability of Al-assisted Alzheimer's detection."
Disclaimer: Curated by HT Syndication.
