MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202541134093 A) filed by Karpagam Academy Of Higher Education; Karpagam Institute Of Technology; K Renuga; and Logeswari V, Coimbatore, Tamil Nadu, on Dec. 31, 2025, for 'ai based cancer detection from medical images.'
Inventor(s) include K Renuga; and Logeswari V.
The application for the patent was published on Jan. 9, under issue no. 02/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to an artificial intelligence (AI)-driven system for automated cancer detection, classification, and segmentation using medical imaging data. Cancer diagnosis traditionally relies on manual interpretation of radiological and pathological images, a process that is time-consuming, subjective, and prone to human error. The disclosed system addresses these limitations by integrating advanced deep learning models with multimodal imaging analysis to provide accurate, consistent, and real-time diagnostic assistance. The invention comprises an image acquisition module that supports multiple medical imaging modalities, including MRI, CT scans, X-ray, ultrasound, and histopathological slides. A preprocessing unit enhances the input data through normalization, noise reduction, contrast adjustment, and region-of-interest extraction, ensuring optimal feature quality for model training and inference. A deep neural network architecture, incorporating convolutional neural networks (CNNs), transformer-based backbones, and optional graph neural networks, is employed to extract multi-scale and modality-specific features from the processed images. A multimodal fusion framework integrates heterogeneous features using attention mechanisms or feature alignment strategies, enabling robust tumor detection across diverse imaging sources. The system includes a segmentation module designed using encoder-decoder architectures such as UNet or Attention UNet, which generates pixel-level tumor masks for accurate delineation of malignant regions. A classification module further analyzes the extracted features to distinguish between benign and malignant tissues and to provide associated confidence scores. To support clinical interpretability, the system incorporates an explainable AI (XAI) module that generates saliency maps, heatmaps, and feature highlighting to visually justify the model's predictions. A cloud-enabled deployment infrastructure facilitates real-time inference, remote diagnostics, and seamless integration with existing hospital information systems. The invention additionally employs federated learning techniques to enable collaborative model training without compromising patient privacy, thereby ensuring compliance with secure data-handling standards. Overall, the proposed AI-based cancer detection system provides a reliable, scalable, and efficient solution for enhancing the accuracy and accessibility of modern oncology diagnostics."
Disclaimer: Curated by HT Syndication.
