MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641110681 A) filed by New Horizon College Of Engineering on September 15, 2026, for An Adaptive Secure Image Steganography System Using Cnn-Based Roi Detection And Q-Learning.
Inventors include Akshatha P S; Sushma P; Sonia Maria Dsouza; Sanjeev Kaulgud; and Arpitha T.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: The present invention relates to an adaptive secure image steganography system and method for concealing confidential information within digital images using artificial intelligence. The system comprises an input cover image module (101), an image preprocessing module (102), a CNN-based Region of Interest detection engine (103), an ROI feature extraction module (104), a secret message input module (105), a Fernet encryption engine (106), a Q-learning decision engine (107), an adaptive embedding module (108), a stego image generation module (109), and a secure stego output module (110). The CNN identifies suitable embedding regions, while the secret message is encrypted using the Fernet encryption engine before embedding. The Q-learning decision engine dynamically determines optimal embedding strategies using reward-based policy learning to maximize image quality and embedding security. The adaptive embedding module inserts encrypted information into selected regions of the cover image, producing a secure stego image with enhanced imperceptibility, robustness, and resistance against steganalysis. The invention provides intelligent, adaptive, and secure information hiding suitable for privacy-preserving communication and secure multimedia applications.
Disclaimer: Curated by HT Syndication.
