MUMBAI, India, Feb. 13 -- Intellectual Property India has published a patent application (202541124562 A) filed by Jerusalem Engineering College, Chennai, Tamil Nadu, on Dec. 10, 2025, for 'aut signsiamnet: offline signature verification using siamese network and resnet 18.'
Inventor(s) include Dr. D. Divya; Dr. A. Velmurugan; N. Javed; and R. Chitra Devi.
The application for the patent was published on Feb. 13, under issue no. 07/2026.
According to the abstract released by the Intellectual Property India: "Signature verification plays a crucial role in personal authentication and security systems. Traditional approaches rely on dynamic features such as stroke order, pressure, and speed, requiring specialized hardware like stylus-enabled devices or biometric pads. However, these methods are impractical for verifying offline signatures obtained from scanned or printed documents. This study presents an offline signature verification system that employs deep learning, utilizing ResNetl8 for feature extraction and a Siamese neural network to learn signature similarity. The model is trained using publicly available datasets, where signature pairs are analyzed to classify them as genuine or forged. Our system eliminates the dependency on specialized hardware, making it accessible through simple image capture using cameras or scanners. Additionally, we design a lightweight demo interface to facilitate testing and usability. Experimental evaluation demonstrates promising accuracy and robustness, highlighting the potential of the proposed method for practical biometric authentication applications."
Disclaimer: Curated by HT Syndication.
