MUMBAI, India, Sept. 7 -- Intellectual Property India has published a patent application (202611084257 A) filed by Manipal University Jaipur on July 09, 2026, for An End-To-End License Plate Character Recognition System Under Complex Weather Conditions Using A Dual-Module Conditional Generative Adversarial Network (dmc-Gan).
Inventor includes Dr. Divya Sharma.
The application for the patent was published on September 04, 2026, under issue no. 36/2026.
Abstract: The present invention relates to an end-to-end license plate character recognition system under complex weather conditions using a dual-module conditional generative adversarial network (DMC-GAN). The system comprising a license plate detection module configured to detect and crop a license plate region from an input vehicle image using a YOLOv7-based detection algorithm; a license plate restoration module operatively coupled thereto, comprising a Dual-Module Conditional Generative Adversarial Network (DMC-GAN) incorporating deformable convolution layers and MobileViT skip connections to restore weather-degraded license plate images; and a character recognition module comprising a PaddleOCR engine configured to recognize alphanumeric characters from the restored license plate image, thereby enabling accurate recognition under rain, fog, haze, snow, low illumination, and motion blur with reduced computational overhead. The proposed system, tested on a database containing 1,001 images from the real world. The invention enhances the image restoration and character recognition output of the weather degraded license plate images.
Disclaimer: Curated by HT Syndication.
