MUMBAI, India, Oct. 12 -- Intellectual Property India has published a patent application (202641119751 A) filed by Pvp Siddhartha Institute Of Technology on October 06, 2026, for Implementation Of Adaptive Recursive Gaussian Filters For Biomedical Image Processing Using Zynq Fpga Devices.

Inventors include Dr. Ch. Gangadhar; Dr. T. Maha Lakshmi; Dr. Mohammad Habibulla; D. Praveena Bai; and Ch. Sri Lakshmi.

The application for the patent was published on October 09, 2026, under issue no. 41/2026.

Abstract: ABSTRACT: Title: Implementation of Adaptive Recursive Gaussian filters for Biomedical Image processing using Zynq FPGA Devices The present invention relates to an adaptive recursive Gaussian filtering system and method for real-time biomedical image processing using a Zynq FPGA device. The system dynamically determines recursive Gaussian filter coefficients according to spatially varying noise characteristics, tissue information, local gradients, and prediction errors. A robust noise estimation module determines local noise using wavelet residuals and/or Median Absolute Deviation, while a tissue-aware prior is obtained from CT Hounsfield- unit statistics, MRI tissue-class probabilities, or segmentation information. An adaptive Kalman filtering module estimates a smoothed image state and performs innovation-based covariance matching. A Recursive Least Squares module with a variable forgetting factor further refines the recursive coefficients according to local prediction error. A stability enforcement module constrains recursive poles within a stable region and regularizes coefficients toward a nominal Gaussian response. The stabilized coefficients are applied to forward and backward recursive filtering stages along horizontal and vertical directions. FPGA-based pipelining, parallel arithmetic processing, buffering, and state management provide high-throughput and low-latency operation, enabling adaptive noise reduction while preserving diagnostically relevant biomedical structures.

Disclaimer: Curated by HT Syndication.