Presentation Details
AI-Driven Automated Insurance Claim Damage Assessment System
ID：34
View Protection：ATTENDEE
Updated Time：2026-07-22 16:09:16 Hits：30
Online
Start Time：2026-07-30 17:35（Asia/Kolkata）
Duration：15min
Session：S6 Artificial Intelligence Use Cases »
Presentation File
Attachment File
Tips: Only the registered participant can access the file. Please sign in first.
Abstract
Manual inspection is still the primary method used to assess vehicle insurance claims; therefore, there are longer timeframes and higher operational costs associated with manual assessments. There are inconsistencies with evaluations and vulnerability to fraud within the existing assessment method. This study presents a framework called VisionClaimNet, which provides an AI-based automated vehicle damage assessment procedure by employing deep learning and computer vision for the intelligent processing of vehicle insurance claims. The proposed VisionClaimNet model is structured as a multi-stage architecture composed of the following four components: image preprocessing, YOLOv8 for damage identification/detection, CNN for damage severity classification, and regression analysis for estimating repair costs. Through the use of an interactive web interface that uploads images of vehicles, VisionClaimNet processes these images and will identify/mark/pinpoint areas of damage on the vehicle, classify the damage based on severity, estimate the cost to repair, and generate a preliminary claims report. In order to enhance the robustness and generalizability of the VisionClaimNet model, various image normalization and augmentation techniques were used. In addition, in order to detect suspicious claims, an anomaly-aware analysis technique has been developed to identify unusual patterns based on the potential for fraud associated with each claim.Experimental results indicated that VisionClaimNet produced high detection rates, low latency of inference, and consistent performance with regards to assessing damage across multiple categories of damage. VisionClaimNet automates much of the claims process, thus reducing manual effort, accelerating settlement timeframes, and providing efficiencies that will positively impact current operations of modern insurance companies.
Keywords
Damage Detection and Assessment, Automated Insurance Claims, Deep Learning; YOLOv8, Computer Vision, CNN, Damage Severity Classification, Cost Estimation, Fraud Detection, and Intelligent Insurance Analytics.
Speaker
Shaik Imran
Student Santhiram Engineering College ；Department of Data Science N.Venkatesh Naik
Professor Santhiram Engineering CollegeSubmission Author
All Comments
Important Date
-
Conference Date
Jul 30
2026
toAug 01
2026
-
Jul 28 2026
Draft paper submission deadline
-
Jul 28 2026
Registration deadline
Sponsored By
The United Societies of Science
Organized By
Kongunadu College of Engineering and Technology
Supported By
IEEE Section
IEEE Madras Section
IEEE Madras Section
Contact Information
Previous Conferences
-
2026-07-30 India Trichy
2026 International Conference on Networks Computers and Communications
Submit Comment