Automated Odometer Reading (OdoCap)

Use case

The cost of a car insurance premium is typically calculated based on the miles driven and that doesn’t always have to work for the best interest of the driver, or the car insurance provider. The issue is that odometer readings are often imprecise and unreliable. Considering how important this piece of information is, Cognexa has set out to fix this problem using computer vision technology.

A large car insurance company has approached us to assess whether we could develop a pay-per-mile system, based on automatic reading of vast amounts of dashboard images (odometer readings) to capture the actual car mileage while incorporating security measures to identify any potential manipulation with the picture (fraud prevention).

Solution

In order to minimize (if not eliminate entirely) the human factor in reading dashboard images taken and uploaded from smartphones, we developed a cognitive system based on a deep neural network that has the following functionality:

• determines the right location of the odometer on the dashboard image
• determines the right number to be read
• automatically reads the mileage
• determines the car make and model

Two weeks into the project, we had developed a viable MVP and only a few weeks later we finalized OdoCap with a reading accuracy of 99.9%. This was pretty exciting given the fact that the images were of different quality, taken in different light conditions and from different angles, which made the process even more complex and interesting. The car make and model accuracy was at 98% (the system is trained on 25 different car makes and models – mostly European).

In case of interest, it takes only several days of deep learning to add additional car makes and models.

Result

It took our development team 10 weeks to create OdoCap. The solution has already proved to be effective by cutting human work-time spent on image evaluation by a staggering 90%. Furthermore, our system is 100% consistent and error-free.

OdoCap can be deployed on-premise or as a cloud solution with a mobile version that is available on request. If you are interested in using OdoCap, contact us.

In the interest of full disclosure…

Our system still experiences some difficulties in recognizing a Bugatti Veyron make and model :-), but that is solely due to not having enough training data. Should you have some Veyron dashboard images, please share them with us so we can fill in the knowledge gaps and improve OdoCap’s performance!

P.S. Would you like to learn more about our other solutions regarding vehicle identity verification, vehicle self-inspection etc. ? Please continue on Vehicle onboarding and management automation subpage or you could try our odometer reading demo online.