OCR Qualities Detector

Qualities detector service for OCR.

Supported qualities:

Blur

This quality check if the image is blur or not.

Key	Description	Type
value	Determine whether the image is blur or not	Boolean
confidence	Confidence score for the given results, whether it predicts "blur" or "not blur"	Float
confidence_blur	This confidence indicates the certainty of the model when it predicts "blur"	Float

Example output


    1
2
3
4
5
6
7

    {
  "blur": {
    "confidence": 1.0,
    "confidence_blur": 1.0,
    "value": true
  }
}

Dark

This quality check if the image is dark or not. The image is considered dark by the average pixel value of the image.

Key	Description	Type
value	Determine whether the image is dark or not	Boolean
confidence	Confidence score for the given results, whether it predicts "dark" or "not dark"	Float
confidence_dark	This confidence indicates the certainty of the model when it predicts "dark"	Float

Example output


    1
2
3
4
5
6
7

    {
  "dark": {
    "confidence": 0.9993858337402344,
    "confidence_dark": 0.9993858337402344,
    "value": true
  }
}

Bright

This quality check if the image is bright or not. The bright is determined by the average of the image pixel value.

Key	Description	Type
value	Determine whether the image is bright or not	Boolean
confidence	Confidence score for the given results, whether it predicts "bright" or "not bright"	Float
confidence_bright	This confidence indicates the certainty of the model when it predicts "bright"	Float

Example output


    1
2
3
4
5
6
7

    {
  "bright": {
    "confidence": 0.9999924898147583,
    "confidence_bright": 0.9999924898147583,
    "value": true
  }
}

Flash

This quality check if the image is taken with flash or not. The flash is determined by flash spot on the image.

Key	Description	Type
value	Determine whether the image is flash or not	Boolean
confidence	Confidence score for the given results, whether it predicts "flash" or "not flash"	Float
confidence_flash	This confidence indicates the certainty of the model when it predicts "flash"	Float

Example output


    1
2
3
4
5
6
7

    {
  "flash": {
    "confidence": 0.9999924898147583,
    "confidence_flash": 0.9999924898147583,
    "value": true
  }
}

Crop

This quality check if the document is cropped or not. The crop is determined by the content of the document.

Key	Description	Type
value	Determine whether the document is cropped or not	Boolean
confidence	Confidence score for the given results, whether it predicts "crop" or "not crop"	Float
confidence_crop	This confidence indicates the certainty of the model when it predicts "crop"	Float


    1
2
3
4
5
6
7

    {
  "crop": {
    "confidence": 0.9999924898147583,
    "confidence_crop": 0.9999924898147583,
    "value": true
  }
}

Rotated

This quality check if the document is rotated or not. The rotation is determined by the content of the document.

Key	Description	Type
value	Document rotation value. Available value `0_degree`, `90_degree`, `180_degree`, and `270_degree`	String
confidence	Confidence score for the given results	Float
fixed_image	The fixed image if the image is rotated in `base64`	String

Example output


    1
2
3
4
5
6
7

    {
  "rotated": {
    "confidence": 1.0,
    "fixed_image": "iVBORw0KGgoAAAA...",
    "value": "90_degree"
  }
}

Photocopy

This quality check if the document is photocopy or not. The photocopy is determined by the document color, whether it's colored or black and white.

Key	Description	Type
value	Determine whether the document is photocopy or not	Boolean
confidence	Confidence score for the given results, whether it predicts "photocopy" or "not photocopy"	Float
confidence_photocopy	This confidence indicates the certainty of the model when it predicts "photocopy"	Float


    1
2
3
4
5
6
7

    {
  "photocopy": {
    "confidence": 0.9999924898147583,
    "confidence_photocopy": 0.9999924898147583,
    "value": true
  }
}

Screen / Spoofing

This quality check if the document is taken from screen or not.

Key	Description	Type
value	Determine whether the document is spoofing or not	Boolean
confidence	Confidence score for the given results, whether it predicts "screen" or "not screen"	Float
confidence_screen	This confidence indicates the certainty of the model when it predicts "screen"	Float

Example output


    1
2
3
4
5
6
7

    {
  "screen": {
    "confidence": 0.9999924898147583,
    "confidence_screen": 0.9999924898147583,
    "value": true
  }
}

Document Classifier

This module classify the document type.

Key	Description	Type
value	Document type	String
confidence	Confidence score for the given results	Float

Available document type:

Document	Description
`ktp`	Kartu Tanda Penduduk (Indonesian identity card)
`npwp`	Nomor Pendaftaran Wajib Pajak
`stnk`	Surat Tanda Nomor Kendaraan
`bpkb-1`	Buku Pemilik Kendaraan Bermotor page 1 (Identitas Pemilik)
`bpkb-2`	Buku Pemilik Kendaraan Bermotor page 2 (Identitas Kendaraan)
`bpkb-3`	Buku Pemilik Kendaraan Bermotor page 3 (Dokumen Registrasi Pertama)
`bpkb-4`	Buku Pemilik Kendaraan Bermotor page 4 (Halaman Terakhir)
`passport`	Passport
`kk`	Kartu Keluarga
`singapore-nric`	Singapore National Registration Identity Card
`singapore-work-permit`	Singapore S PASS, Work Permit, and Employment Pass
`singapore-family-pass`	Singapore Dependant Pass and Long-Term Visit Pass

Example output


    1
2
3
4
5
6

    {
  "document": {
    "confidence": 0.9999924898147583,
    "value": "ktp"
  }
}

POST/ocr/v1/qualities

Detect Qualities

Detects the supported qualities from uploaded image.

Required form parameter

Name
image
Type
file (.png, .jpg, .jpeg)
Description
The image file of the document.

Sample Request

POST

/ocr/v1/qualities


    1
2
3
4

      curl -v -L -X POST 'https://api.vision.glair.ai/ocr/v1/qualities' \
    -H "Authorization: Basic $(printf "%s" "USERNAME:PASSWORD" | base64)" \
    -H 'x-api-key: API_KEY' \
    -F 'image=@"/path/to/image.jpg"'

Sample Response


    1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50

    {
  "qualities": {
      "blur": {
          "confidence": 1.0,
          "confidence_blur": 1.0,
          "value": false
      },
      "bright": {
          "confidence": 0.9999924898147583,
          "confidence_bright": 0.9999924898147583,
          "value": false
      },
      "crop": {
          "confidence": 0.9999818801879883,
          "confidence_crop": 0.9999818801879883,
          "value": false
      },
      "dark": {
          "confidence": 0.9993858337402344,
          "confidence_dark": 0.9993858337402344,
          "value": false
      },
      "document": {
          "confidence": 1.0,
          "value": "ktp"
      },
      "flash": {
          "confidence": 0.9991378784179688,
          "confidence_flash": 0.9991378784179688,
          "value": false
      },
      "photocopy": {
          "confidence": 0.9999991655349731,
          "confidence_photocopy": 0.9999991655349731,
          "value": false
      },
      "rotated": {
          "confidence": 1.0,
          "fixed_image": "iVBORw0KGgoAAAA...",
          "value": "0_degree"
      },
      "screen": {
          "confidence": 0.9889932153746486,
          "confidence_screen": 0.9889932153746486,
          "value": false
      }
  },
  "reason": "File successfully read.",
  "status": "SUCCESS"
}

Optional query parameter

Name
qualities
Type
string (supported qualities above)
Description
The quality only you want to detect

Sample Request

POST

/ocr/v1/qualities


    1
2
3
4

    curl -v -L -X POST 'https://api.vision.glair.ai/ocr/v1/qualities?qualities=dark,flash,blur' \
  -H "Authorization: Basic $(printf "%s" "USERNAME:PASSWORD" | base64)" \
  -H 'x-api-key: API_KEY' \
  -F 'image=@"/path/to/image.jpg"'

Sample Response


    1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18

    {
  "qualities": {
      "blur": {
          "confidence": 1.0,
          "value": false
      },
      "dark": {
          "confidence": 0.9993858337402344,
          "value": false
      },
      "flash": {
          "confidence": 0.9991378784179688,
          "value": false
      }
  },
  "reason": "File successfully read.",
  "status": "SUCCESS"
}

Request ID

An associated request identifier is generated for every request made to this endpoint. This value can be found in the response headers under Request-Id

Responses

Various responses for this endpoint, in addition to general responses specified in Errors.

`200 - OK`

Request with readable image

Response


    1
2
3
4
5

    {
  "status": "SUCCESS",
  "reason": "File successfully read.",
  //...,
}

`200 - Unrecognized Document Type`

Request with unrecognized document type

Response


    1
2
3
4
5

    {
  "status": "UNRECOGNIZED_DOCUMENT_TYPE",
  "reason": "Unrecognized document type detected.",
  //...,
}

`200 - Multiple Documents`

Request with multiple documents

Response


    1
2
3
4
5

    {
  "status": "MULTIPLE_DOCUMENTS_DETECTED",
  "reason": "Multiple documents detected.",
  //...,
}

`400 - Bad Request`

Request without form-data image

Response


    1
2
3
4
5

    {
  "status": "FAILED",
  "reason": "Input should be an instance of FileStorage",
  //...,
}

`400 - Bad Request`

Request with non supported file format

Response


    1
2
3
4
5

    {
  "status": "FAILED",
  "reason": "mimetype: application/zip is not supported",
  //...,
}