Degree

Bachelor of Science (Computer Science)

Department

Department of Computer Science

School

School of Mathematics and Computer Science (SMCS)

Advisor

Tasbiha Fatima, Lecturer, Department of Computer Science, Institute of Business Administration, Karachi

Keywords

AI, Vehicle Damage Detection, Machine Learning, Auto-Insurance, Claims

Abstract

The rapid digitization of the insurance sector demands innovative solutions to automate traditionally manual, time-consuming, and error-prone processes. In the domain of auto insurance, claim processing remains a significant bottleneck, causing customer dissatisfaction and operational inefficiencies. This project, IntelliClaims, addresses these challenges by leveraging artificial intelligence to automate vehicle damage detection and repair cost estimation from photographic evidence. Drawing from current advancements in computer vision and machine learning, IntelliClaims integrates a secure, web-based platform with state-of-the-art deep learning models for semantic segmentation and regression analysis. The system streamlines claim submission for insurance employees, automates the assessment of vehicle images using a Mask R-CNN model, and predicts repair costs via advanced regression algorithms. Experimental results demonstrate significant improvements in claim processing speed and accuracy, reducing human intervention while maintaining reliability. The IntelliClaims approach not only expedites workflows but also provides a scalable foundation for future enhancements, including advanced fraud detection and real-time customer feedback mechanisms.

Tools and Technologies Used

TypeScript React, FastAPI, TensorFlow, Postgresql, Google Colab, Github

Methodology

Proposed Approach:

We propose a comprehensive, end-to-end vehicle claim assessment pipeline that automates damage identification and cost estimation using deep learning and modern web technologies. This solution comprises three core modules:

1. Damage and Severity Detection: Powered by instance segmentation models using the Detectron2 framework.

2. Repair Cost Estimation: Implemented via a machine learning regression model based on XGBoost.

3. Integrated Claim Management System: A full-stack application using a React-based frontend and FastAPI backend to manage the complete claims workflow.

Workflow Overview:

The claims workflow begins when employees upload images of a damaged vehicle to the system. These images are automatically processed by the AI-powered backend, which detects damaged parts, classifies the severity of the damage, and estimates repair costs. The system then generates a detailed report and updates the claims dashboard, facilitating faster and more accurate decision-making. This pipeline reduces dependency on manual inspections, significantly accelerates processing time, and enhances accuracy and consistency in claim assessments. Additional system features include:

● End-to-end claim submission and approval tracking.

● Automated damage detection and report generation.

● A centralized dashboard for analytics and claim history.

● Export functionality for AI-generated assessment reports.

Document Type

Restricted Access

Submission Type

BSCS Final Year Project

Download

Share

COinS