MUMBAI, India, June 30 -- Intellectual Property India has published a patent application (202641053821 A) filed by Mepco Schlenk Engineering College on April 28, 2026, for Tri-Channel Feature Fusion Denoising Model For Target Detection.
Inventors include Dr. S. Arivazhagan; Dr. W. Sylvia Lilly Jebarani; T. Pooja; and S. N. Aishwarya.
The application for the patent was published on June 26, 2026, under issue no. 26/2026.
Abstract: ABSTRACT: The invention discloses a tri-channel feature fusion based denoising model for improving target detection in Synthetic Aperture Radar (SAR) images. The proposed model integrates spatial, frequency, and non-local feature representations using convolutional neural networks, wavelet transforms, and non-local means filtering. These features are fused to generate a denoised image with enhanced structural and contextual information. The processed image is further utilized for accurate target detection using deep learning-based detection models. Experimental results demonstrate improved noise suppression, higher structural similarity, and enhanced detection performance compared to conventional methods. The proposed approach is suitable for military surveillance and remote sensing applications.
Disclaimer: Curated by HT Syndication.
