MUMBAI, India, Aug. 17 -- Intellectual Property India has published a patent application (202511003639 A) filed by Lovely Professional University on January 16, 2025, for Hybrid Cnn-Svm System For Handwriting Recognition.
Inventors include Dr. Manikrakhra; Tiyas Sarkar; Aryan Maan; and Baljinderkaur.
The application for the patent was published on August 14, 2026, under issue no. 33/2026.
Abstract: The present invention relates to a handwriting recognition system that integrates a Convolutional Neural Network (CNN) and a Support Vector Machine (SVM) into a hybrid architecture to enhance the accuracy and efficiency of recognizing handwritten text. The system comprises a CNN configured for feature extraction from input images, utilizing multiple layers of convolutions, pooling, and non-linear activations to generate high-level feature representations. An SVM is employed to classify these feature vectors by finding an optimal hyperplane that maximizes the margin between different classes. A data preprocessing module ensures input data is normalized, resized, and noise-reduced for consistent processing by the CNN. The system includes a training and optimization module that employs backpropagation for the CNN and hyperparameter tuning for the SVM, facilitating the iterative adjustment of model parameters. An inference engine is designed for real-time deployment, processing new input data and outputting recognized text. Dated this 28th Day of Dec, 2024 Dr. Monica Gulati Registrar Lovely Professional University
Disclaimer: Curated by HT Syndication.
