Monday, August 19, 2019

AI system to assist with insurance claims management

Example of a Computer Vision Architecture
Figure shows the implementation of a computer-vision-based AI system to assist with insurance claims management.

Images can be captured and uploaded via a mobile device (1) which is then sent for pre-processing (2) to adjust contrast, color and size for noise reduction. It is then put into an image database and training set (3). Here, location information (image metadata) that is captured from the devices can be used for fraud detection by using streaming data from the mobile device. Now, an image processing system, leveraging the ML model (4) and built for comparing images of good and damaged vehicles, assesses the damage. This then helps identify parts required to repair the damaged vehicle and associates any labor costs (5).
This completely automates the process of assessment and makes it easier for the auto damage adjuster to process the claim quickly. However, this wouldn't have been possible if it weren't for machine learning models that were trained to use lots of image data over deep learning systems to help build an end-to-end AI system that leverages computer vision.