Tencent Cloud

Recent Pages

Blind Watermarking

Last updated: 2024-02-18 10:36:36
Overview
The blind watermark feature allows you to add an invisible watermark to an image, which does not affect the look and feel and retains the originality of the image. If you suspect that your image has been stolen, you can extract the invisible watermark through blind watermark extraction to prove your ownership of the image.
You can call this API to:
Add or extract a blind watermark to or from an image during upload.
Add or extract a blind watermark to or from an image stored in COS.
Add a blind watermark to an image during download.
CI introduces three blind watermarking types: semi-blind watermarking, perfectly blind watermarking, and text blind watermarking.
Blind Watermarking Type
Feature
Applicable Scenario
Semi-blind watermarking (type 1)
Has strong anti-theft effects but requires the input image for watermark extraction
Images smaller than 640x640
Perfectly blind watermarking (type 2)
Is easy to extract and requires only the watermarked image for watermark extraction
Batch adding and verification
Text blind watermarking (type 3)
Adds text to the image
Adding terminal information
Blind watermarking is a paid service, which needs to be activated on the bucket's configuration page. For detailed pricing, see Image Processing Fees.
Note: 
Currently, you cannot add a blind watermark to animated images such as GIF.
Both the width and height of the image watermark must not be greater than 1/8 of the input image.
You need to choose a white watermark with a black background for the effect of blind watermarking.
Text watermarks can contain digits and letters.
Blind watermarking protects your images against various theft attacks such as cropping, smudging, and color change. The anti-theft effect is affected by the size of the input image and the attack intensity. For details, please submit a ticket.
Blind watermarking is supported in all regions of CI. For more information, see Regions and Domain Names.
Use Cases
Authentication and accountability
By adding semi-blind watermarks to your images, you can claim your images when they are stolen and extract blind watermarks using the corresponding input images to prove that you own the images.
Checking duplicates upon upload
Sometimes, other users might upload duplicate images (such as real estate images, car images, and product images). To solve this problem, you can perform perfectly blind watermark extraction before the upload, and if you can extract a blind watermark from an image, it has been uploaded before. In this case, you can perform corresponding operations, such as telling the user not to upload duplicate images. If an image does not contain a blind watermark, you can add one to it to prevent duplicate uploads.
Avoiding image leakage
For internal images, you can add information about the requester to the images by using text blind watermarks. In this way, when an image is leaked, you can extract the blind watermark and get information about the leaker.
Adding Blind Watermarks
Request 1: adding a blind watermark upon upload
The request is the same as that in a PUT Object call, except that you need to add the image processing parameter Pic-Operations to the request header and use blind watermarking parameters.
Sample request
PUT /<ObjectKey> HTTP/1.1
Host: <BucketName-APPID>.cos.<Region>.myqcloud.com
Date: GMT Date
Authorization: Auth String
Pic-Operations: <PicOperations>
﻿
[Object Content]
Note: 
Authorization: Auth String (See Request Signature for details.)
When this feature is used by a sub-account, relevant permissions must be granted as instructed in Authorization Granularity Details.
Request headers
This API uses Common Request Headers.
Add the image processing parameter Pic-Operations (a JSON string) to the request header. Its parameters are as follows:
Parameter
Description
Type
Required
is_pic_info
Whether to return the input image information 
0 (default): no 
1: yes
Int
No
rules
Processing rules (up to five rules are supported). Each rule corresponds to one processing result. If this parameter is not specified, images will not be processed.
Array
No
Parameters of rules (a JSON array) are as follows:
Parameter
Description
Type
Required
bucket
Name of the destination bucket to store the results, formatted as BucketName-APPID. If this parameter is not specified, the results will be stored in the current bucket.
String
No
fileid
Storage path and name of the output file. 
Naming rules are as follows (assume the input file is /p1/test1.jpg):
1. A value starting with a slash (/) indicates an absolute path. For example, if fileid is set to /p2/test2.jpg, the test2.jpg file will be stored in the p2 folder. 
2. A value not starting with a slash indicates a relative path. For example, if fileid is set to p2/test2.jpg, a folder named p2 will be created in the p1 folder, and the test2.jpg file will be stored in the p2 folder.
3.  Note: Do not end the value with a slash; otherwise, an empty filename will be generated.
String
Yes
rule
Processing parameters. For more information, see the image processing API. To process an image by using a specified style, the value must start with style/, with the style name followed. For example, if the style name is test, the value of rule should be style/test.
String
Yes
To use blind watermarking, add the watermarking parameter watermark to rule as follows:
watermark/3/type/<type>/image/<imageUrl>/text/<text>/level/<level>
The parameters are as follows:
Parameter
Description
Type
Required
type
Blind watermarking type. Valid values: 1 (semi-blind watermarking), 2 (perfectly blind watermarking), 3 (text blind watermarking)
Int
Yes
image
URL of the blind image watermark, which must be URL-safe Base64 encoded. This parameter is required if type is set to 1 or 2, and does not take effect if type is set to 3. The image watermark must:
1. Be stored in the same bucket as the input image. 
2. Have a URL that starts with http://. Note that "http://" cannot be omitted or changed to "https://". For example, the following watermark URLs are invalid: 
examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
https://examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
String
No
text
Text watermark, which must be URL-safe Base64-encoded. This parameter is required if type is set to 3 and does not take effect if type is set to 1 or 2.
String
No
level
Perfectly blind watermarking level, which only takes effect when type is set to 2. Valid values: 1 (default), 2, 3. The larger the value, the better the watermarking effects.
Int
No
Request parameters
This API has no request parameter.
Request body
The request body of this API is the content of the input image.
Response
Response headers
This API only returns Common Response Headers.
Response body
Parameters in the response body are as follows:
Parameter
Description
Type
UploadResult
Upload results
Container
Content of UploadResult: 
Parameter
Description
Type
OriginalInfo
Information about the input image
Container
ProcessResults
Processing results
Container
Content of OriginalInfo:
Parameter
Description
Type
Key
Name of the input image
String
Location
Image location
String
ImageInfo
Information about the input image
Container
Content of ImageInfo:
Parameter
Description
Type
Format
Image format
String
Width
Image width
Int
Height
Image height
Int
Quality
Image quality
Int
Ave
Image average hue
String
Orientation
Image rotation angle
Int
Content of ProcessResults:
Parameter
Description
Type
Object
Processing results of each image
Container
Content of Object:
Parameter
Description
Type
Key
Filename
String
Location
Image location
String
Format
Image format
String
Width
Image width
Int
Height
Image height
Int
Size
Image size
Int
Quality
Image quality
Int
Example
Request
PUT /exampleobject HTTP/1.1
Host: examplebucket-1250000000.cos.ap-chengdu.myqcloud.com
Date: Tue, 03 Apr 2018 09:06:15 GMT
Authorization:XXXXXXXXXXXX
﻿
Pic-Operations:
{
    "is_pic_info": 1,
    "rules": [{
        "fileid": "exampleobject",
        "rule": "watermark/3/type/1/image/aHR0cDovL2V4YW1wbGVzLTEyNTEwMDAwMDQucGljc2gubXlxY2xvdWQuY29tL3NodWl5aW4uanBn"
    }]
}
Content-Length: 64
﻿
[Object Content]
Response
HTTP/1.1 200 OK
Content-Type: application/xml
Content-Length: 645
Date: Tue, 03 Apr 2018 09:06:16 GMT
Status: 200 OK
x-cos-request-id: NWFjMzQ0MDZfOTBmYTUwXzZkZV8z****
﻿
<UploadResult>
          <OriginalInfo>
              <Key>exampleobject</Key>
              <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject</Location>
              <ImageInfo>
                   <Format>JPEG</Format>
                   <Width>640</Width>
                   <Height>427</Height>
                   <Quality>100</Quality>
                   <Ave>0xa18454</Ave>
                   <Orientation>1</Orientation>
              </ImageInfo>
          </OriginalInfo>
          <ProcessResults>
              <Object>
                   <Key>exampleobject</Key>
                   <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject</Location>
                   <Format>png</Format>
                   <Width>640</Width>
                   <Height>427</Height>
                   <Size>463092</Size>
                   <Quality>100</Quality>
              </Object>
          </ProcessResults>
</UploadResult>
Note: 
[Object Content] in the request body is the binary string of the input image, and imageUrl in rule in the request header Pic-Operations is the URL of the image watermark.
Location in ProcessResults in the response body is the URL of the image with a blind watermark.
Request 2: adding a blind watermark upon download
You can add a blind watermark upon download in the same way as adding a regular watermark, except that you need to use the watermark parameter after the image URL.
Sample request
GET /<ObjectKey>?watermark/3/type/<type>/image/<imageUrl>/text/<text> HTTP/1.1
Host: <BucketName-APPID>.cos.<Region>.myqcloud.com
Date: <GMT Date>
Authorization: <Auth String>
Request parameters
The request parameters are as follows:
Parameter
Description
Type
Required
ObjectKey
Object name, such as folder/sample.jpg.
String
Yes
type
Blind watermarking type. Valid values: 1 (semi-blind watermarking), 2 (perfectly blind watermarking), 3 (text blind watermarking)
Int
Yes
image
URL of the blind image watermark, which must be URL-safe Base64 encoded. This parameter is required if type is set to 1 or 2, and does not take effect if type is set to 3. The image watermark must:
1. Be stored in the same bucket as the input image. 
2.  Have a URL that starts with http://. Note that "http://" cannot be omitted or changed to "https://". For example, the following watermark URLs are invalid: 
examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
https://examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
String
No
text
Text watermark, which must be URL-safe Base64-encoded. This parameter is required if type is set to 3 and does not take effect if type is set to 1 or 2.
String
No
level
Perfectly blind watermarking level, which only takes effect when type is set to 2. Valid values: 1 (default), 2, 3. The larger the value, the better the watermarking effects.
Int
No
Example
https://examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/sample.jpeg?watermark/3/type/3/text/dGVuY2VudCBjbG91ZA==
Request 3: adding a blind watermark to an image stored in COS
You can call this API to add a blind watermark to an image stored in COS and save the result to COS.
Sample request
POST /<ObjectKey>?image_process HTTP/1.1
Host: <BucketName-APPID>.cos.<Region>.myqcloud.com
Date: GMT Date
Authorization: Auth String
Pic-Operations: <PicOperations>
Note: 
Authorization: Auth String (See Request Signature for details.)
When this feature is used by a sub-account, relevant permissions must be granted as instructed in Authorization Granularity Details.
Request headers
This API uses Common Request Headers.
Add the image processing parameter Pic-Operations (a JSON string) to the request header. Its parameters are as follows:
Parameter
Description
Type
Required
is_pic_info
Whether to return the input image information 0 (default): no 1: yes
Int
No
rules
Processing rules (up to five rules are supported). Each rule corresponds to one processing result. If this parameter is not specified, images will not be processed.
Array
No
Parameters of rules (a JSON array) are as follows:
Parameter
Description
Type
Required
bucket
Name of the destination bucket to store the results, formatted as BucketName-APPID. If this parameter is not specified, the results will be stored in the current bucket.
String
No
fileid
Storage path and name of the output file.
Naming rules are as follows (assume the input file is /p1/test1.jpg):
1. A value starting with a slash (/) indicates an absolute path. For example, if fileid is set to /p2/test2.jpg, the test2.jpg file will be stored in the p2 folder. 
2.  A value not starting with a slash indicates a relative path. For example, if fileid is set to p2/test2.jpg, a folder named p2 will be created in the p1 folder, and the test2.jpg file will be stored in the p2 folder.
3.  Note: Do not end the value with a slash; otherwise, an empty filename will be generated.
String
Yes
rule
Processing parameters. For more information, see the image processing API. To process an image by using a specified style, the value must start with style/, with the style name followed. For example, if the style name is test, the value of rule should be style/test.
String
Yes
To use blind watermarking, add the watermarking parameter watermark to rule as follows:
watermark/3/type/<type>/image/<imageUrl>/text/<text>/level/<level>
The parameters are as follows:
Parameter
Description
Type
Required
type
Blind watermarking type. Valid values: 1 (semi-blind watermarking), 2 (perfectly blind watermarking), 3 (text blind watermarking)
Int
Yes
image
URL of the blind image watermark, which must be URL-safe Base64-encoded. This parameter is required if type is set to 1 or 2, and does not take effect if type is set to 3. The image watermark must:
1. Be stored in the same bucket as the input image. 
2. Have a URL that starts with http://. Note that "http://" cannot be omitted or changed to "https://". For example, the following watermark URLs are invalid: 
examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
https://examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
String
No
text
Text watermark, which must be URL-safe Base64-encoded. This parameter is required if type is set to 3 and does not take effect if type is set to 1 or 2.
String
No
level
Perfectly blind watermarking level, which only takes effect when type is set to 2. Valid values: 1 (default), 2, 3. The larger the value, the better the watermarking effects.
Int
No
Request parameters
This API has no request parameter.
Request body
This API does not have a request body.
Response
Response headers
This API only returns Common Response Headers.
Response body
Parameters in the response body are as follows:
Parameter
Description
Type
UploadResult
Upload results
Container
Content of UploadResult: 
Parameter
Description
Type
OriginalInfo
Information about the input image
Container
ProcessResults
Processing results
Container
Content of OriginalInfo: 
Parameter
Description
Type
Key
Name of the input image
String
location
Location of the input image
String
ImageInfo
Information about the input image
Container
Content of ImageInfo: 
Parameter
Description
Type
Format
Image format
String
Width
Image width
Int
Height
Image height
Int
Quality
Image quality
Int
Ave
Image average hue
String
Orientation
Image rotation angle
Int
Content of ProcessResults:
Parameter
Description
Type
Object
Processing results of each image
Container
Content of Object:
Parameter
Description
Type
Key
Filename
String
location
Location of the input image
String
Format
Image format
String
Width
Image width
Int
Height
Image height
Int
Size
Image size
Int
Quality
Image quality
Int
Example
Request
POST /exampleobject?image_process HTTP/1.1
Host: examplebucket-1250000000.cos.ap-chengdu.myqcloud.com
Date: Tue, 03 Apr 2018 09:06:15 GMT
Authorization:XXXXXXXXXXXX
﻿
Pic-Operations:
{
    "is_pic_info": 1,
    "rules": [{
        "fileid": "exampleobject",
        "rule": "watermark/3/type/1/image/aHR0cDovL2V4YW1wbGVzLTEyNTEwMDAwMDQucGljc2gubXlxY2xvdWQuY29tL3NodWl5aW4uanBn"
    }]
}
Content-Length: 64
Response
HTTP/1.1 200 OK
Content-Type: application/xml
Content-Length: 645
Date: Tue, 03 Apr 2018 09:06:16 GMT
Status: 200 OK
x-cos-request-id: NWFjMzQ0MDZfOTBmYTUwXzZkZV8z****
﻿
<UploadResult>
          <OriginalInfo>
              <Key>exampleobject</Key>
              <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject</Location>
              <ImageInfo>
                   <Format>JPEG</Format>
                   <Width>640</Width>
                   <Height>427</Height>
                   <Quality>100</Quality>
                   <Ave>0xa18454</Ave>
                   <Orientation>1</Orientation>
              </ImageInfo>
          </OriginalInfo>
          <ProcessResults>
              <Object>
                   <Key>exampleobject</Key>
                   <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject</Location>
                   <Format>png</Format>
                   <Width>640</Width>
                   <Height>427</Height>
                   <Size>463092</Size>
                   <Quality>100</Quality>
              </Object>
          </ProcessResults>
</UploadResult>
Extracting Blind Watermarks
Example 1: Extracting upon upload
Sample request
The request for uploading an image with the blind watermark extracted is the same as that in a PUT Object call, except that you need to add the image processing parameter Pic-Operations to the request header and use the watermark extraction parameter watermark/4.
PUT /<ObjectKey> HTTP/1.1
Host: <BucketName-APPID>.cos.<Region>.myqcloud.com
Date: GMT Date
Authorization:XXXXXXXXXXXX
Pic-Operations: <PicOperations>
﻿
[Object Content]
Note: 
Authorization: Auth String (See Request Signature for details.)
When this feature is used by a sub-account, relevant permissions must be granted as instructed in Authorization Granularity Details.
Request headers
This API uses Common Request Headers.
Add the image processing parameter Pic-Operations (a JSON string) to the request header. Its parameters are as follows:
Parameter
Description
Type
Required
is_pic_info
Whether to return the input image information 0 (default): no 1: yes
Int
No
rules
Processing rules (up to five rules are supported). Each rule corresponds to one processing result. If this parameter is not specified, images will not be processed.
Array
No
Parameters of rules (a JSON array) are as follows:
Parameter
Description
Type
Required
bucket
Name of the destination bucket to store the results, formatted as BucketName-APPID. If this parameter is not specified, the results will be stored in the current bucket.
String
No
fileid
Storage path and name of the output file. 
Naming rules are as follows (assume the input file is /p1/test1.jpg):
1. A value starting with a slash (/) indicates an absolute path. For example, if fileid is set to /p2/test2.jpg, the test2.jpg file will be stored in the p2 folder. 
2.  A value not starting with a slash indicates a relative path. For example, if fileid is set to p2/test2.jpg, a folder named p2 will be created in the p1 folder, and the test2.jpg file will be stored in the p2 folder.
3.  Note: Do not end the value with a slash; otherwise, an empty filename will be generated.
String
Yes
rule
Processing parameters. For more information, see the image processing API. To process an image by using a specified style, the value must start with style/, with the style name followed. For example, if the style name is test, the value of rule should be style/test.
String
Yes
To extract a blind watermark, you need to add the watermark parameter to rule as follows:
watermark/4/type/<type>/image/<imageUrl>
The parameters are as follows:
Parameter
Description
Type
Required
type
Blind watermarking type. Valid values: 1 (semi-blind watermarking), 2 (perfectly blind watermarking), 3 (text blind watermarking). The value must be the same as that used when the blind watermark was added.
Int
Yes
image
Image URL, which should be specified according to the value of type. If the value of type is: 
1, image is required and is the URL of the input image. 
2, image is required and is the URL of the image watermark. 
3, image does not need to be specified and will not take effect. 
The value of image must be URL-safe Base64-encoded, and the image must: 1. Be stored in the same bucket as the image with the blind watermark. 2. Have a URL that starts with http://. Note that "http://" cannot be omitted or changed to "https://". For example, the following watermark URLs are invalid: 
examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
https://examplebucket-1250000000.cos.ap-shanghai.myqcloud.com/watermark.png
String
No
Request parameters
This API has no request parameter.
Request body
The request body of the API is the binary string of the image.
Response
Response headers
This API only returns Common Response Headers.
Response body
WatermarkStatus will be added to the UploadResult.ProcessResults.Object node only if type is set to 2 in the request body of the blind watermark extraction request.
Parameter
Parent Node
Description
Type
WatermarkStatus
Object
This parameter is returned if type is set to 2 to indicate the confidence of the watermark extracted. Value range: 0–100. 
If the value is greater than 75, there is a blind watermark. 
If the value falls in the range of 60−75, there is a suspected blind watermark. 
If the value is smaller than 60, no blind watermark is extracted.
Int
Example
Request
PUT /exampleobject1 HTTP/1.1
Host: examplebucket-1250000000.cos.ap-chengdu.myqcloud.com
Date: Tue, 03 Apr 2018 09:06:15 GMT
Authorization:XXXXXXXXXXXX
﻿
Pic-Operations:
{
    "is_pic_info": 1,
    "rules": [{
        "fileid": "exampleobject2",
        "rule": "watermark/4/type/2/image/aHR0cDovL2V4YW1wbGVzLTEyNTEwMDAwMDQucGljc2gubXlxY2xvdWQuY29tL2ZpbGVuYW1lLmpwZWc="
    }]
}
Content-Length: 64
﻿
[Object Content]
Response
HTTP/1.1 200 OK
Content-Type: application/xml
Content-Length: 645
Date: Tue, 03 Apr 2018 09:06:16 GMT
Status: 200 OK
x-cos-request-id: NWFjMzQ0MDZfOTBmYTUwXzZkZV8z****
﻿
<UploadResult>
      <OriginalInfo>
          <Key>exampleobject1</Key>
          <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject1</Location>
          <ImageInfo>
                <Format>JPEG</Format>
                <Width>640</Width>
                <Height>427</Height>
                <Quality>100</Quality>
                <Ave>0xa18454</Ave>
                <Orientation>1</Orientation>
                <FrameCount>1</FrameCount>
          </ImageInfo>
      </OriginalInfo>
      <ProcessResults>
          <Object>
                <Key>exampleobject2</Key>
                <Location>examplebucket-1250000000.cos.ap-chengdu.myqcloud.com/exampleobject2</Location>
                <Format>png</Format>
                <Width>640</Width>
                <Height>427</Height>
                <Size>463092</Size>
                <Quality>100</Quality>
                <FrameCount>1</FrameCount>
                <WatermarkStatus>86</WatermarkStatus>
          </Object>
      </ProcessResults>
</UploadResult>
Note: 
If type is set to 1 in the request, imageUrl in rule in Pic-Operations is the URL of the input image without a blind watermark added. If type is set to 2 in the request, imageUrl is the URL of the image watermark.
url in ProcessResults in the response body is the URL of the image watermark extracted.
Example 2: Extracting from an image stored in COS
The request for extracting a blind watermark from an image stored in COS is the same as that used to Process In-Cloud Data, except that you need to add the image processing parameter Pic-Operations to the request header and use the watermark extraction parameter watermark/4.
​	
Contact Us
Contact our sales team or business advisors to help your business.
Technical Support
Open a ticket if you're looking for further assistance. Our Ticket is 7x24 avaliable.
7x24 Phone Support
Blind Watermarking Type	Feature	Applicable Scenario
Semi-blind watermarking (type 1)	Has strong anti-theft effects but requires the input image for watermark extraction	Images smaller than 640x640
Perfectly blind watermarking (type 2)	Is easy to extract and requires only the watermarked image for watermark extraction	Batch adding and verification
Text blind watermarking (type 3)	Adds text to the image	Adding terminal information
Parameter	Description	Type	Required
is_pic_info	Whether to return the input image information `0` (default): no `1`: yes	Int	No
rules	Processing rules (up to five rules are supported). Each rule corresponds to one processing result. If this parameter is not specified, images will not be processed.	Array	No
Parameter	Parent Node	Description	Type
WatermarkStatus	Object	This parameter is returned if `type` is set to `2` to indicate the confidence of the watermark extracted. Value range: 0–100. If the value is greater than 75, there is a blind watermark. If the value falls in the range of 60−75, there is a suspected blind watermark. If the value is smaller than 60, no blind watermark is extracted.	Int
tencent cloud

Sign Up

Log in

Recent Pages

Blind Watermarking

Overview

Use Cases

Authentication and accountability

Checking duplicates upon upload

Avoiding image leakage

Adding Blind Watermarks

Request 1: adding a blind watermark upon upload

Sample request

Request headers

Request parameters

Request body

Response

Response headers

Response body

Example

Request

Response

Request 2: adding a blind watermark upon download

Sample request

Request parameters

Example

Request 3: adding a blind watermark to an image stored in COS

Sample request

Request headers

Request parameters

Request body

Response

Response headers

Response body

Example

Request

Response

Extracting Blind Watermarks

Example 1: Extracting upon upload

Sample request

Request headers

Request parameters

Request body

Response

Response headers

Response body

Example

Request

Response

Example 2: Extracting from an image stored in COS

Was this page helpful?

Was this page helpful?
