/face/anti_hack

1.API Specification
2.Request Params
3.Response Params
4.Error Codes
5.Sample

1.API Specification

This API is used to check if it has hack behavior in the uploaded image.

Image Spec

1. Format: JPG(JPEG), PNG
2. Between 128*128 and 6000*6000 (pixels)
3. File size: no larger than 5 MB

Request Method

POST

Request URL

https://cloudapi.accuauth.in/face/anti_hack

Debugging Tool

2.Request Params

2.1 Request Header

Parameter	Type	Required	Description
X-DF-API-ID	string	Yes	For API credentials，Please visit API Request
X-DF-API-SECRET	string	Yes	For API credentials，Please visit API Request

2.2 Request Body

Fields	Type	Required	Description
file	file	See the note below	Image file，binary data of the image
image_base64	string	See the note below	Base64 encoded data of image
return_extra	int	No	Whether to return extra information; 0: No, 1: Yes, default is 0

Either parameter file or image_base64 must be provided as a request parameter.

You MUST add image stream to multipart/form-data section of POST message when using file parameter.

3.Response Params

Parameter	Type	Description
request_id	string	Unique id of each request
status	string	Response status, `OK` if request success, other if failed; See Error Codes for details
score	float	Hack score [0.0-1.0], the greater of the value the greater probablitity of non-human
extra	object	Extra information, see extra for detail

NOTE: The recommended hackness score threshold value is 0.98, which is set based on training data. If hackness score is below 0.98, it would be a real person.

3.1 Elements in extra

Parameter	Type	Description
mask_proba	float	Probability of the face is masked, 0.5 is the recommended threshold
quality_score	float	Quality score, 0.5 is the recommended threshold
quality_too_low	bool	Whether the image quality is too low
blur_or_dark	bool	Whether the image is blur or dark
has_no_face	bool	Whether the image has a face
face_too_small	bool	Whether the face is too small
image_too_small	bool	Whether the image is too small
masked	bool	Whether the face is masked
face_quality_low	bool	Whether the face quality is too low
face_quality_score	bool	Face quality score, 0.5 is the recommended threshold
is_live	bool	Whether the face is live
live_proba	bool	Liveness score, 0.5 is the recommended threshold
eye_closed	bool	Whether the eyes are closed
eye_score	float	Probability of the eyes being open, 0.5 is the default threshold

Response Syntax

{
  "request_id": "TID8bf47ab6eda64476973cc5f5b6ebf57e",
  "status": "OK",
  "score":"0.13119",
  "extra": {
    "mask_proba": 0.001,
    "quality_score": 0.961,
    "quality_too_low": false,
    "blur_or_dark": false,
    "has_no_face": false,
    "face_too_small": false,
    "image_too_small": false,
    "masked": false,
    "face_quality_low": false,
    "face_quality_score": 0.9981434181918459,
    "is_live": true,
    "live_proba": 0.9967548623045266,
    "eye_closed": false,
    "eye_score": 0.9786062829240756
  }
}

4.Error Codes

Error code	Status value	Description
`400`	INVALID_ARGUMENT	The request parameter error
`400`	DETETION_FAILED	Detect image failed
`400`	DOWNLOAD_ERROR	Download internet image failed
`401`	UNAUTHORIZED	Unauthorize or authorize failed
`401`	KEY_EXPIRED	Your API ID has expired
`403`	NO_PERMISSION	You are not authorized to use this API
`403`	OUT_OF_QUOTA	API calls quota exceeded
`403`	RATE_LIMIT_EXCEEDED	The invoke number is out of limitation
`404`	NOT_FOUND	The request path is not found
`500`	INTERNAL_ERROR	Server processes failed

Reponse Syntax:

{
  "status": "INVALID_ARGUMENT",
  "reason": "must specify 'file' or 'image_base64' argument",
  "request_id": "TID8bf47ab6eda64476973cc5f5b6ebf57e"
}

Image Anti-hack