I am a horrible driver, when I first hit my back bumper, it took me a long time to get the evaluation of the damage. I stayed in line, went to several different places, just want to get a basic idea about how bad is the damage. Finally, I didn't fix it because it is not bad enough for me to pay the money. I want something that can give me sense about the damage level of my car.
What it does
It is a website that the user can upload the image of the damaged part of the car, then the website can scale the level of damage and give advise. The insurance company can use this is a quantifier in the contract of compensation.
How we built it
We use the API give by Clarifai, it helps us to build up our own model of quantifying the damage. Then we employ that into the website which the domain name given by domain.com.
Challenges we ran into
Categorizing the damage and train the model with the correct data set
Accomplishments that we're proud of
We built up our own model by machine learning. We feed the model data about relationship between damage and level than let it learn.
What we learned
Knowledge about machine learning, how to build a website, how to connect front end and back end
What's next for HowBad
The model for HowBad can be used not only for evaluation car damage. We can feed it with different sets of data and let it learn.