Pixurance — Quick Car Insurance
A React Native mobile app for car accident insurance claim assessment — using YOLO v11 computer vision model and Roboflow platform for real-time damage detection.
The Project
Pixurance addresses the pain point of slow, manual car insurance claim assessment. By deploying a YOLO v11 computer vision model on a React Native mobile app, field assessors and policyholders can capture photos of vehicle damage and receive instant preliminary damage classification.
Roboflow handles the model training pipeline, dataset management, and edge deployment — while the React Native frontend provides a production-grade mobile experience for claim intake and assessor workflows.
Technical Highlights
YOLO v11 fine-tuned on a vehicle damage dataset covering 5 damage classes, deployed via Roboflow.
Roboflow for dataset versioning, model training pipeline, and edge inference API.
React Native for cross-platform mobile (iOS + Android) — real-time image capture and CV inference.
Quantized model optimised for mobile edge inference, enabling near real-time damage classification.
Building a computer vision or mobile AI system?