Inspiration
During our research, we found that a lot of people face difficulties in claiming their motor damage in a timely manner. It would be beneficial to have a system for the customer to make claims faster and for insurance providers to set aside the claim cost accurately. Thus, we decided to create an app where the customer can submit claims and receive an estimate repair cost by simply take a picture of the damage. To avoid panics after an accident, we also provide instruction for after-accident care. We want to help facilitate a more efficient claim process system.
What it does
Our AI-based app can instruct the customer what to do after an accident happens, identify the motor damage parts and severity with our well-trained ML model, provide an estimate repair and claim cost and send a complete report to the insurance provider.
How I built it
We used both TensorFlow and TensorFlow Lite to train on thousands of collision images we scraped from the internet. We classified them accurately and trained the model to identify them. Then we integrate the trained model into an Android app with Android Studio.
Challenges I ran into
We have tried different methods to perfect the model. None of us have experience with Android development so we learned along the way
Accomplishments that I'm proud of
Didn't give up on the last day and finished the project
What I learned
Tensorflow Lite, Android Studio and teamwork
What's next for Project Winston - Your Car Insurance Companion
Upgrade the UX and add more functionalities
Built With
- android
- android-studio
- flutter
- java
- keras
- python
- tensorflow
- tensorflowlite
- waston
Log in or sign up for Devpost to join the conversation.