Damage Detector Model for car manufacturing companies

The client is one of the largest financial services companies in the world. Their objective was to develop world class AI & ML models for their various products for which machines are required to be trained using high quality and unbiased datasets achieved through human evaluation process.

Engagement

GlobalLogic has been engaged with the client for 4 years to create a damage detector model by identifying, classifying and labeling the damages caused to several types of cars like Hatchbacks, Sedans & SUVs. We evaluated multiple car manufacturing companies like Toyota, KIA, Ford, GM etc., and labeled more than 20 million images of cars, trucks and sedan models.

Process

As part of the labeling process, GlobalLogic identified the damages occurred to the cars which are sent for insurance claims. Once the damages are identified, they are classified into categories like Dents, Scratches, Crush etc. These damages are labeled on the specific part of the car, and this way we train the model to identify damages through the damage detector.

Outcome

The project eased the documentation process for the claims and no-claims procedures for insurance companies. This engagement was also critical in developing vehicle damage detectors for Sedans and Pickup trucks, which allowed us to move into the commercial phase too.

Let’s start engineering impact together.

GlobalLogic provides unique experience and expertise at the intersection of data, design, and engineering.

Ontology – key Enabler of Next-Gen Technologies

Smartphone on Wheels

Damage Detector Model for car manufacturing companies