For this we used a CNN model, VGG16[1]. We used our own dataset that we had prepared and trained the VGG16 model ourselves. Our study consisted of 4 main questions:
- Is it a car?
- Is there any damage on the car?
- In which part of the car is the damage?
- What is the level of damage?
We prepared and trained our own datasets within these 4 questions.
Data Sets | Training | Validation |
---|---|---|
Is it a car? | 920 |
230 |
Is there any damage on the car? | 1840 |
460 |
In which any part of the car is the damage? | 976 |
171 |
What is the level of damage? | 979 |
171 |
We used our original dataset, which consists of a total of 5,757 photographs.
Data Sets | Training | Validation |
---|---|---|
Car | 920 |
230 |
The result we will get here is only querying whether there is a car or not.
Data Sets | Training | Validation |
---|---|---|
Damaged | 920 |
230 |
Undamaged | 920 |
230 |
We have done this training in order to determine whether there is a similar damage at this stage and to continue with the other stages according to the result.
Data Sets | Training | Validation |
---|---|---|
Front | 418 |
73 |
Rear | 287 |
50 |
Side | 271 |
48 |
our aim here is to determine which part of the car the damaged area in the photo belongs to. We trained our parameters as front, back and side.
Data Sets | Training | Validation |
---|---|---|
Minor | 278 |
48 |
Moderate | 315 |
55 |
Severe | 386 |
68 |
Now that we know the damage is and where it is, it's up to us to separate the level of damage we set ourselves.
Training | Accuracy Rating |
---|---|
Is it a car? | %98 |
Is there any damage on the car? | %90 |
In which any part of the car is the damage? | %70 |
What is the level of damage? | %66 |