To automate the human inspection and cost estimation procedure in the auto insurance sector, we created an iOS app for vehicle damage recognition and dent depth profiling. The software finds dents and scratches, measures their height and width, and offers a depth profile to calculate the time and cost of repairs. Three components of our AI engine operate with a tolerance of 1 cm. The app has helped the client increase productivity by 500% with no user biasness. The technology stack includes Python, Swift, custom DL models, OpenCV, and Yolo.
Business Challenge:
The global auto insurance industry is around $800 billion. One primary use case is the manual inspection of vehicle dents and then estimation for its repair. This process is human intensive and varies with every evaluator. The client to automate this whole procedure with minimum input from human.
Solution:
We have developed an IOS app that automates the dent inspection and cost estimation process with minimum human intervention. The user takes pictures of a vehicle with our app. The app detects any scratches or dents visible in the picture. Furthermore, it not only gives the height and width of the dent, but also the depth profile of the dent. This data along with the type of panel on which the dent is present is used to estimate the time and price estimate for repairing the dent.
AI Engine Results:
There are three AI modules. First module detects the dents and scratches. The second module measures the height and width within tolerance of 1 cm, and the third module provides the depth profile within tolerance of 1 cm.
Results:
An IOS app that automatically scans the vehicle and gives realistic price and time estimates.
Client Success:
The client increased the productivity by 500% along with zero user biasness.
Technology Stack:
Language: Python, Swift
Technologies: Custom DL Models, OpenCV, Yolo