Overview
Introduction
Our Vehicle Damage Inspection API provides the identification of the car part, the type of damage and extent of damage. We also provide whether the damage requires a repair or a replacement.
Link here: https://docs.google.com/spreadsheets/d/1gS-srFSAjgUceYMVHUeJCVjAARAvQ0UWrws-IkhfQyA/edit#gid=0
Video url
https://www.youtube.com/watch?v=yHykpvtU8ME
Pricing
The trial access to the API costs $499 / month for 1000 inspection (1 inspection = 8-10 images). We do not provide a free trial.
Features
Quick features we'd like to highlight -
1. The API is intelligent enough to identify even tiny scratches and scuffs as in the video.
2. The API is intelligent enough to not to confuse reflection of rain droplets etc on the front windshield etc as damaged or broken glass.
FAQs
1. How can I test out the API?
Feel free to ask us for a live demo over a call. Feel free to share 5-10 images with us to share the results of the model with you. Once you’re confident with the api, you can get started with the $499 monthly plan. You can cancel anytime and just pay as you go.
2. What is the typical response time?
30s - 1min / car
3. What kind of limitations does your product have in terms of image quality, size, lighting?
The API can handle pretty low-res images we've seen in the past, but our recommended image size is >1000x1000
4. Does your system allow training? ie if we provide it with photos of the same vehicle undamaged would it perform better.
All your data is being used to re-train your model to improve accuracy. Yes that would contribute to our positive sample data space.
5. My volume is significantly greater than 1000 inspections. Can you provide me with better pricing?
Ofcourse, please get in touch with our sales team for economical pricing for higher volumes.
6. Will the API work for trucks and other vehicles?
Yes, the API is Vehicle agnostic (any car, truck, etc.)
7. Is there a recommendation for how to capture the images for the model to perform the best?
The following angles would work best to capture images. Side frame shots work great. Close ups are not preferred.
8. Is the model accurate in determining severity at two levels for scratches, dents etc.?
Yes, the model even recommends whether a repair or replacement of the car part is required.
9. Can the API recognize a car's brand and model?
Not currently available since this information is already available with the insurance provider.
Do we need to retrain the model with images of a new car?
No, the model has already been trained with thousands of vehicle images across the world. This is already pre-built in the model.
Sample API Response
Sample API response for the image below.
("area" below refers to the % area of the damage = (area of damage) / (total area of the part)
{
"status": "success",
"message": "success",
"result": [
{
"status": "Success",
"data": {
"quoteId": "Q-2yASpHEMW",
"quoteType": "lease",
"segmentImages": {
"rawImages": [
"https://nanonetscustomerdata.s3-us-west-2.amazonaws.com/wfn/images/kegwtwky_2-6.jpg "
],
"annotatedImages": [
"https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
],
"imageDetails": [
{
"_id": "5ede122df7028169305fb9d5",
"raw": "https://nanonetscustomerdata.s3-us-west-2.amazonaws.com/wfn/images/kegwtwky_2-6.jpg ",
"annotated": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg ",
"filename": null,
"error": null
}
]
},
"segmentationEstimate": {
"damages": [
{
"leftfrontdoor": [
{
"name": "scratch",
"area": 48.04,
"url": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
},
{
"name": "scratch",
"area": 0.09,
"url": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
},
{
"name": "d1",
"area": 0.06,
"url": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
}
],
"damageCode": "Dent1"
},
{
"doorhandle": [
{
"name": "clean",
"area": 100,
"url": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
}
],
"damageCode": "Clean"
},
{
"leftorvm": [
{
"name": "clean",
"area": 100,
"url": "https://customerstaging.nanonets.com/wfn-processed-images/annotated_data/Q-2yASpHEMW_1591611941392_annotated_1591611948455.jpg "
}
],
"damageCode": "Clean"
},
{
"leftrunningboard": [
{
"name": "scratch",
"area": 0.36,
"url": "
