MUMBAI, India, Aug. 12 -- Intellectual Property India has published a patent application (202641086520 A) filed by Nandha Engineering College on July 15, 2026, for Image-Based Malware Detection Using Deep Learning For Cybersecurity Applications.
Inventors include A Arputhamary; and K Subhashini.
The application for the patent was published on August 07, 2026, under issue no. 32/2026.
Abstract: This invention introduces an efficient, deep learning system that uses static binary visualization to identify and categorize multi-family malware. By transforming raw executable bytes into normalized 2D grayscale images, the framework conducts safe structural evaluations without executing the threat. It replaces traditional signature- based matching with an automated, custom Sequential CNN designed to extract structural textures and spatial features. The model accurately classifies malware variants across 25 distinct strains with a proven 96% accuracy rate. Connected to a Flask web dashboard backend, the system ensures real-time security scanning, risk analysis, and alert generation, making it highly .effective for enterprise defense and network forensic
Disclaimer: Curated by HT Syndication.
