MUMBAI, India, June 22 -- Intellectual Property India has published a patent application (202641069920 A) filed by Sasi Institute Of Technology & Engineering; V Srinivas; M Satya Srinivas; S Siva Ramaraja; and Mpavanisasi. Ac. In on June 04, 2026, for Deep Learning Approach For Vehicle Damage Analysis And Repair Cost Estimation.
Inventors include V Srinivas; M Satya Srinivas; S Siva Ramaraja; and Mpavanisasi. Ac. In.
The application for the patent was published on June 12, 2026, under issue no. 24/2026.
Abstract: The present invention relates to a deep learning-based vehicle damage detection and repair cost estimation system for automated vehicle inspection and analysis. The system is designed to identify and classify vehicle damages from uploaded images and provide intelligent repair cost estimation using artificial intelligence techniques. The proposed system integrates an object detection model based on YOLOv5 for detecting various types of vehicle damages such as dents, scratches, cracks, broken lamps, and shattered glass. The system further includes an image pre-processing module for enhancing input image quality through resizing, normalization, and augmentation techniques to improve detection accuracy. The invention additionally incorporates a damage severity analysis module for classifying detected damages into severity levels including minor, moderate, and severe. An AI-based repair cost estimation module, utilizing a language model-based reasoning approach, is employed to estimate repair costs based on detected damage type, severity, and vehicle-specific details such as model, brand, and manufacturing year. The system is implemented using a web-based architecture comprising a frontend interface and a Flask-based backend for real-time processing and user interaction. The output is presented through a user-friendly interface displaying detected damage regions, severity classification, and estimated repair cost along with explanatory insights.Experimental evaluation demonstrates that the proposed system achieves high detection accuracy with approximately 95% precision, 92% recall, and strong mean Average Precision (mAP) values, ensuring reliable performance in real-world conditions. The invention provides an efficient, scalable, and automated solution for vehicle damage assessment, reducing manual inspection efforts, improving insurance claim processing efficiency, and enabling accurate and rapid repair cost estimation.
Disclaimer: Curated by HT Syndication.
