Skip to content
/ VAMRR-Vehicle-Detection Public
forked from nirbhayph/VAMRR-Vehicle-Detection

Deep learning based vehicle damage detection solution. This repository is linked to the web application which is integrated with the mentioned system.

autodetect.ml
3 stars 3 forks Branches Tags Activity
Star
Notifications You must be signed in to change notification settings

DhirenSc/VAMRR-Vehicle-Detection

BranchesTags
 
 

Repository files navigation

Deep Learning based Vehicle Damage Detection

Demo : https://autodetect.ml

This is a Deep learning based vehicle damage identification solution. This repository is linked to the React based web application which is integrated with the mentioned system.

This is a proof of concept for damage detection of vehicles. Labels that can be detected : Front Rear Door Fender Major-Damage Minor-Damage Scratches

The React based web application involves uploading your media file(s), waiting for them to get processed and finally seeing your predicted outputs.

API Documentation

Authorization

Every user is identified by their authorization key. All API requests require authentication from a specific user. For example, seeking sessions for a particular user.

Thus, to authenticate requests, pass your access key in the following format:

data {
  key:"<SPECIFY AUTHORIZATION_KEY HERE>",
  ...
}

Creating and using existing sessions

Every time a user wishes to make a new request, they need to specify the sessionId which can take either of the two values:

data {
  session:"new",
  ...
}

Here, new indicates a new sessionId to be created.

If the user wishes to not create a new session, the API allows the option to merge new requests with previous request results by specifying an older sessionId.

data {
  session:"9943213153939",
  ...
}

Specifying API Endpoints

The url attribute can take the following endpoints:

url:'https://autodetect.ml/api/request/',
data: {
  ...
}

url:'https://autodetect.ml/api/lookup/',
data: {
  ...
}

url:'https://autodetect.ml/api/seek/',
data: {
  ...
}

Supplying media input to the API /request POST

It is important that the media URL(s) are specified through an array only even if there is only one URL whose output needs to be processed.

.js example

var url_array = [
    'https://www.autoauctionmall.com/learning-center/contents/uploads/2015/10/damaged-car.jpg?w=100&h=200&k=300',
    'https://www.autoauctionmall.com/learning-center/contents/uploads/2015/10/damaged-car.jpg?w=100&h=200&k=300'
    ];
data: {
  url:url_array,
  ...
}

POST

All API requests (/request, /lookup, /seek) must have data passed via a POST request ONLY.

type:'POST',
data: {
  ...
}

API /request POST example (Generate JSON output for a request on a new/existing sessionId)

.js example

var url_array = [
    'https://www.autoauctionmall.com/learning-center/contents/uploads/2015/10/damaged-car.jpg?w=100&h=200&k=300',
    'https://www.autoauctionmall.com/learning-center/contents/uploads/2015/10/damaged-car.jpg?w=100&h=200&k=300',
    'https://www.autoauctionmall.com/learning-center/contents/uploads/2015/10/damaged-car.jpg?w=100&h=200&k=300'
    ];

    $(document).ready(function(){
      $.ajax({
      url:'https://autodetect.ml/api/request/',
      type:'POST',
      data:{
        url:url_array,
        key:"<SPECIFY AUTHORIZATION_KEY HERE>",
        session:"5147839406218"
      },
      success:function(result){
      console.log(result);
      }
    });
});

Sample Output

[
    {
        "sessionId": "4370958592436",
        "fileName": "4679fa55e8c79524212c9b60c8ed59de.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/4679fa55e8c79524212c9b60c8ed59de.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4679fa55e8c79524212c9b60c8ed59de.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    },
    {
        "sessionId": "4370958592436",
        "fileName": "7d8ccdbf059c96f9bfccf46786ed137d.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/7d8ccdbf059c96f9bfccf46786ed137d.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/7d8ccdbf059c96f9bfccf46786ed137d.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    },
    {
        "sessionId": "4370958592436",
        "fileName": "94a87f6d2921081b8acd67fe35f3d88c.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/94a87f6d2921081b8acd67fe35f3d88c.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/94a87f6d2921081b8acd67fe35f3d88c.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    }
]

Output explanation for /request POST

sessionId: The sessionId of the API request (A new one is generated if new keyword is specified during /request) fileName: Filename for the .json generated originalImageLink: The link for the original image link supplied during API /request outputImageLink: The output image link pointing to the labeled image visualizationLink: The link to view the entire session via a browser marker: The label specification for the generated output image, along with color coding for each label textOutputLink: The link to the text output specifying processing details for a /request made

API /lookup POST example (Retrieve existing sessionId's data)

.js example

$(document).ready(function(){
      $.ajax({
      url:'https://autodetect.ml/api/lookup/',
      type:'POST',
      data:{
        key:"<SPECIFY AUTHORIZATION_KEY HERE>",
        session:"5147839406218"
      },
      success:function(result){
      console.log(result);
      }
    });
});

Sample Output

[
    {
        "sessionId": "4370958592436",
        "fileName": "4679fa55e8c79524212c9b60c8ed59de.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/4679fa55e8c79524212c9b60c8ed59de.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4679fa55e8c79524212c9b60c8ed59de.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    },
    {
        "sessionId": "4370958592436",
        "fileName": "7d8ccdbf059c96f9bfccf46786ed137d.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/7d8ccdbf059c96f9bfccf46786ed137d.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/7d8ccdbf059c96f9bfccf46786ed137d.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    },
    {
        "sessionId": "4370958592436",
        "fileName": "94a87f6d2921081b8acd67fe35f3d88c.json",
        "originalImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/processed\/94a87f6d2921081b8acd67fe35f3d88c.jpeg",
        "outputImageLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/94a87f6d2921081b8acd67fe35f3d88c.jpeg",
        "visualizationLink": "https:\/\/autodetect.ml\/p.php?p=4370958592436",
        "marker": "{\"label\": \"major-damage\", \"color_rgb_code\": [95.25, 190.5, 127]}, {\"label\": \"front\", \"color_rgb_code\": [254.0, 254.0, 254]}, {\"label\": \"fender\", \"color_rgb_code\": [158.75, 63.5, 127]}, {\"label\": \"minor-damage\", \"color_rgb_code\": [127.0, 254.0, 254]}",
        "textOutputLink": "https:\/\/autodetect.ml\/file_handler\/uploads\/4370958592436\/out\/4370958592436.txt"
    }
]

API /seek POST example (Retrieving existing sessionIds)

Users can retrieve all sessionIds associated with a key.

.js example

$(document).ready(function(){
      $.ajax({
      url:'https://autodetect.ml/api/seek/',
      type:'POST',
      data:{
        key:"<SPECIFY AUTHORIZATION_KEY HERE>"
      },
      success:function(result){
      console.log(result);
      }
    });
});

Sample Output

[
    {
        "sessionId": "11299504729248",
        "lastModifiedTime": "2019-03-29 05:28:17.884912910"
    },
    {
        "sessionId": "5147839406218",
        "lastModifiedTime": "2019-03-29 05:29:07.881948821"
    },
    {
        "sessionId": "6112784572874",
        "lastModifiedTime": "2019-03-29 04:40:23.874150987"
    }
]

Error Messages

If you encounter any of the following messages, you may have not specified the url/data attributes in the valid format.

Invalid Authorization Key

Session Does Not Exist

Post data not found!

No sessions found

Redundant Process Request

Unauthorized Access

Invalid Media Format (.png, .jpeg, .jpg only)*

Invalid Format for Data Provided, (There may possibly be a problem with the Media URL)

Please check your post data!

Invalid Header Contents

Media URLs not passed!

NOTE: If you encounter any other problems, feel free to open an issue.

Authors: (@nirbhayph), (@dhirensc)

VAMRR Technologies Pvt. Ltd.

About

Deep learning based vehicle damage detection solution. This repository is linked to the web application which is integrated with the mentioned system.

autodetect.ml

Topics

car front reactjs vehicle darknet door darknet-image-classification darkflow fender rear vehicles-insurance damage-detection vehicle-damage-detection minor-damage major-damage

Resources

Readme
Activity

Stars

3 stars

Watchers

1 watching

Forks

1 fork
Report repository

Releases

No releases published

Packages

No packages published

Languages

  • HTML 55.8%
  • JavaScript 44.2%