Fraud Detection and Prevention of Motor Insurance using Computer Vision

Technologies has been chosen to Design, Architect, Build, Test, Deploy, Enhance and Maintain the application

• Use FNOL Images/Videos of damaged parts in case of motor accidents
• Compare images received from garages with images of damaged parts with FNOL images
• Map out differences if any (visually, with coordinates) and alert Insurers representative

Software stack – Java 14, Python, MongoDB, OpenCV, Yolov7, Detectron

Cloud Environment – AWS/Azure/Google (TBD)
Radical new approach data, training and usage of CNNs & RNNs have evoked interest from Insurers from Middle East, South and East Africa, South East Asia and India in terms of participating in a Beta.