MUMBAI, India, July 30 -- Intellectual Property India has published a patent application (202641089042 A) filed by Easwari Engineering College; and Srm Institute Of Science And Technology, Ramapuram Campus on July 22, 2026, for Task-Conditioned Feature Modulation System With Fpga Acceleration For Real Time Underground Mining Safety Monitoring.
Inventors include Aneeka Yasmeen R; Charumathi G; Felcita Pavithra F; and Dr. Jessintha Devaraj.
The application for the patent was published on July 24, 2026, under issue no. 30/2026.
Abstract: The present invention relates to a hardware-software co-designed system for task-conditioned object detection in underground mining safety monitoring. A shared neural-network backbone is conditioned using Feature-wise Linear Modulation (FiLM), wherein task-specific scale and shift coefficients are applied to intermediate feature tensors, enabling a single neural-network model to support multiple safety monitoring tasks without model reloading or retraining. A dedicated hardware accelerator implemented on a reconfigurable logic device performs FiLM operations through a parallel fixed-point pipeline, retrieving task-specific coefficients from on-chip memory via a memory-mapped register interface. Runtime task switching is achieved by updating a task register without modifying network parameters or reconfiguring hardware. Detection results are processed using a priority-tier selection mechanism to report object classes relevant to the active task, including protective-equipment verification, zone intrusion detection, machinery-proximity monitoring, conveyor-belt monitoring, material accumulation detection, and personnel headcount for real-time edge deployment. Total number of words in abstract:148
Disclaimer: Curated by HT Syndication.
