Machine Vision for Automotive Damage Cost Estimation

• 

  Entity-Relationship Diagram

Inspiration

The current processing time of automotive insurance claims can result in stressful customers in this fast-paced society. Our main focus is customers' and service providers' satisfaction by estimating damage cost to reduce claim processing time.

What It Does

This application enables customers to click and send picture(s) of the damaged part of the automotive to assess repairing/replacement cost by implementing machine vision. There are four categories of the classification, scratch, dent, damaged, and undamaged.

How We Built It

We have extended current State Farm application to apply transfer learning from Google's Inception V3 pre-trained model that will effectively estimate your unique situation.

Challenges We Ran Into

The biggest challenge was to develop this model in a short span of time and integrating the android application coded in java with Tensor flow typically in python.

Accomplishments That We're Proud Of

We are able to develop an important feature of this model, i.e.estimating the repair/replacement cost based on the damage by utilizing the picture of the damaged part of the vehicle.

What We Learned

During the development of this model, we learned general insurance process and we also learned that PHP is better for online database integration with Cordova Android development.

What's Next For Machine Vision For Automotive Damage Cost Estimation

Though we have achieved a satisfactory result; however, we need to have several improvements to make this model better, for example an effort can be made to improve accuracy of this model, automatically classify damaged part, develop a module for an accurate assessment of the possible damage underneath vehicle body, and also location and other data can be utilized to estimate the damage cost or other possible results (e.g. whether the vehicle is repairable or not).

Built With

• apache
• css
• html5
• javascript
• sql
• tensorflow