MUMBAI, India, Sept. 28 -- Intellectual Property India has published a patent application (202641113030 A) filed by Vallurupalli Nageswara Rao Vignana Jyothi Institute Of Engineering And Technology on September 21, 2026, for Deep Learning-Based Secure Communication System Integrating Hybrid Cryptography, Randomized Lsb Steganography, And Image Classification.
Inventor includes Dr. Harshavardhan Awari.
The application for the patent was published on September 25, 2026, under issue no. 39/2026.
Abstract: ABSTRACT [0022] The invention provides a dual-layer secure steganographic system that integrates hybrid cryptography with image steganography to safeguard secret messages. A sender supplies a plaintext message and a cover image through a graphical interface. The message is first encrypted by a hybrid scheme that applies a Playfair substitution cipher followed by a columnar transposition cipher using a shared symmetric key. The resulting ciphertext is converted to binary and embedded into the least significant bits of randomly selected pixel positions within the cover image. The chosen embedding locations are recorded in a position file that is itself encrypted. To further conceal the presence of hidden data, approximately fifty visually identical duplicate images are generated and transmitted together with the single stego-image. On the receiving side, a ResNet50 feature extractor followed by K-means clustering automatically isolates the true stego-image from the set of duplicates. The encrypted position file is then decrypted and used to extract the embedded bits, which are reassembled and decrypted with the same hybrid cipher to recover the original message. Experimental evaluation demonstrates high imperceptibility, with a mean squared error of 0.0175, a peak signal-to-noise ratio of 65.69 dB, and a structural similarity index of 0.9994, together with strong resistance to statistical and brute-force analysis.
Disclaimer: Curated by HT Syndication.
