MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641110796 A) filed by Andhra University on September 15, 2026, for Automated Brain Tumour Segmentation From Mr Images Using A Modified Dual-Network Architecture With Multi-Level Attention.
Inventors include Kommu Sangeet Kumar; Prof. B Biswal; and Prof. P. Sridevi.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: ABSTRACT: Title: Automated Brain Tumour Segmentation from MR Images Using a Modified Dual-Network Architecture with Multi-Level Attention This invention presents a hybrid ANN-SFOA model for the high-precision prediction of surface air temperature and precipitation, addressing the challenges of the climate crisis. The system integrates Artificial Neural Networks (ANN) with a novel Starfish Optimisation Algorithm (SFOA), which mimics the biological searching and regenerative behaviors of starfish. By employing a multi-armed search mechanism and a regenerative tactic, the SFOA optimizes the weights and biases of the ANN, effectively overcoming the limitations of local optima and slow convergence found in traditional models. The model processes seven key meteorological input features, including cloud cover, humidity, and global radiation. Experimental results indicate that the ANN-SFOA model achieves superior accuracy with an R^2 value exceeding 0.96. This invention provides a robust and efficient tool for hydrological modeling, agricultural planning, and ecological management, significantly enhancing the reliability of weather forecasting in the face of global warming.
Disclaimer: Curated by HT Syndication.
