MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641126906 A) filed by Koneru Lakshmaiah Education Foundation; Eda Bhagyalakshmi; and A K Velmurugan on October 24, 2026, for Integrated Adaptive Medical Image Segmentation Using Dynamic Learning And Anatomical Priors..
Inventors include Koneru Lakshmaiah Education Foundation; Eda Bhagyalakshmi; and A K Velmurugan.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: The present invention provides an integrated adaptive medical image segmentation system using dynamic learning and anatomical priors. The system receives medical images from one or more imaging modalities and processes the images through a computational architecture comprising feature extraction, successive categorical feature fusion, dynamic weighted adversarial learning, hierarchical category consistency, and spatially regularized anatomical prior mechanisms. The extracted image features are progressively combined to improve the representation of relevant anatomical structures. A dynamic weighting mechanism adjusts the contribution of learning components according to image characteristics and variations in data distributions, while adversarial learning supports adaptation across different imaging domains. A hierarchical category consistency mechanism promotes semantic agreement among segmented anatomical regions. The system further incorporates statistical anatomical priors and spatial regularization into the segmentation objective to encourage spatially coherent predictions consistent with learned anatomical information. The segmentation model is optimized using a combined learning objective that integrates image-based segmentation information, adaptive feature learning, category consistency, and anatomical prior constraints. An iterative feedback mechanism updates model parameters to improve segmentation under varying imaging conditions. The invention provides an integrated computational framework for adaptive, anatomically informed, and spatially consistent medical image segmentation.
Disclaimer: Curated by HT Syndication.
