Choose where you’d like to start

 

Detect Face

Overview

The zoho.ai.detectFace task detects all the faces in the image and returns the bounding box coordinates of the faces detected. Additionally, the emotion, gender, and pose attributes of the face can also be detected.

For example, using this task, you can predict if a seminar was engaging based on the audience's facial expressions. Learn more.

Note: 
  • The prediction results may not be accurate, which is the case with any AI prediction. However, we are working on improving this.
  • The prediction results are dynamic. The same script may produce different outcomes based on the machine's knowledge.

Syntax

<response>  =  zoho.ai.detectFace(<input_image>,  <is_dominant_face>,  <attribute_list>);

where:

ParamsData typeDescription

<response>

 KEY-VALUE

 Specifies the predicted face coordinates, emotion, gender, and position of all faces in the   image.

<input_image>

 File

 Specifies the file object that contains the image.

 Note: 
  • The file can only be fetched from the cloud using the invokeUrl task.
  • This task cannot be applied directly to the Zoho Creator image or file upload field.
  • The file can be an image of .png, .jpg, or .jpeg format.
  • The maximum allowed file size is 5MB.

<is_dominant_face>

 Boolean  (true/false)

 Specifies the condition to return the data of prominent face or all faces in the   image. 

 true - Returns bounding box of prominent face.

 false - Returns bounding boxes of all faces in the image.

 

<attribute_list>

(optional)

 List

 Specifies the list of values to return. The supported values are:

  • gender - Includes gender of the face (Male / Female) in the returned response.
  • emotion - Emotion of the face (Angry / Fear / Happy / Neutral / Sad / Surprise)
  • pose -Returns angle/pose of the face (Yaw/Pitch/ Roll)

Example 1: Detect faces in an image

The following script detects the face in the image and returns its coordinates.

image_file  =  invokeurl
    [
         url : "https://c0.wallpaperflare.com/preview/52/764/33/twin-boys.jpg"
      ];

 response  =   zoho.ai.detectFace(image_file, false);

Where:

"https://c0.wallpaperflare.com/preview/52/764/33/twin-boys.jpg"
is the TEXT that represents the URL of the image that needs to be analyzed.
response
is the KEY-VALUE response that represents the objects detected from the image file and their coordinates.
image_file
is the FILE that represents the image from which the faces will be detected.
false
is the BOOLEAN value that specifies the task will return all the faces in the image.
 

Example 2: Detect additional attributes like gender, emotion and pose.

The following script detects the faces in the image and returns its response with face coordinates, gender, emotion, and pose.

image_file  =  invokeurl
     [
         url :"https://c0.wallpaperflare.com/preview/52/764/33/twin-boys.jpg"
     ];
​
     attribute_list  =  list();
     attribute_list.add("gender");
     attribute_list.add("emotion");
     attribute_list.add("pose");

     response  =  zoho.ai.detectFace(image_file, false, attribute_list);

Where:

attribute_list
is the LIST of values that need to be returned. Click here to learn more about list function.

Example 3: Apply zoho.ai.detectface task on an image uploaded in a Zoho Creator field

Note: This example is only applicable for Zoho Creator.

The below code downloads the image from the field through API and returns its response with zoho.ai.detectFace task. The zoho.ai.detectFace task cannot be used with Zoho Creator’s form fields. Before that, you need to establish a connection within the application. In order to download the uploaded file in the form, you need download file API to use it with the zoho.ai.detectFace task.

 // Replace <account_owner_name> with  the  user name of the Creator accounts owner.
 // Replace <app_link_name> the link name of the target application.
 // Replace <report_link_name> with  the link name of the target report.
 // Replace <record_ID > with the ID of the record from which you want to download the file.
 // Replace <field_link_name> with the link name of the target file upload or image field.

 image_file  =  invokeurl
     [
     url: "https://creator.zoho.com/api/v2/<account_owner_name>/<app_link_name>/report/<report_link_name>/<record_ID>/<field_link_name>/download"
     type : GET
     connection: "creator_oauth_connection"
     ];
 
  attribute_list  =  List();
  attribute_list.add("gender");
  attribute_list.add("emotion");
  attribute_list.add("pose");
 
  response  =  zoho.ai.detectFace(image_file,  false,  attribute_list);

Where:

"https://creator.zoho.com/api/v2/<account_owner_name>/
<app_link_name>/report/<report_link_name>/<record_id>
/<field_link_name>/download</field_link_name
</record_id></report_link_name></app_link_name></account_owner_name>"
is the TEXT that represents the URL specified in the Zoho Creator to download the required image through Zoho Creator - Download File API.
creator_oauth_connection
is the TEXT that represents the name of the connection, created using the default service Zoho OAuth. While creating the connection, include scopes as mentioned in the Zoho Creator - Download File API.
 

Scenario 1: Fetch required data from the response

Consider an image where you need to know the total count of people. This can be done by detecting the number of faces in the image. The zoho.ai.detectFace task stores the number of faces found data in the response as numFacesFound. This particular data can be fetched using the get() function. Refer to the following script for better understanding.

 image_file  =  invokeurl
     [
        url : "https://ylpapp.isb.edu/user/Passport_photo_samples/Sample-9---wrong.jpg"
     ];
 // Create a variable to store the response.
 response  =  zoho.ai.detectFace(image_file,  false);
 // Fetch the "numFacesFound" data from response using get() function.
 number_of_faces  =  response.get("numFacesFound");

Scenario 2: Predict, whether the program was engaging based on the facial expression of the audience

Consider an image with a group of people sitting in an auditorium. By analyzing the facial emotions of the people in the picture, we can conclude whether the program is engaging or not.

If the number of happy faces exceeds 50% of the total face count, we can conclude that the program is engaging, otherwise it is not engaging.

The following script fulfils the above-said scenario.

 image_file  =  invokeurl
     [
      url : "https://media.istockphoto.com/photos/diverse-human-faces-picture-id1187245319?k=20&m=1187245319&s=612x612&w=0&h=byaOZWGwYLZI3YoGagQEiFds05DHjmcRfSynTeTrPe4="
      type : GET
     ];
 // Create a variable -face_data to store the data as list by using get() and toList() function.
 face_data  =   zoho.ai.detectFace(image_file, false, {"emotion"}).get("data").toList();
  count  =  0;
 // Create a variable - total_face to stores the total face count in the image using size() function.
 total_face  =  face_data.size(); 

  for each  index  face  in  face_data
  {
     detect_emotion  =  face_data.get(face).get("emotion").get("prediction");

     if(detect_emotion  ==  "happy")
     {
         count  =  count + 1;
     }
  }

  if(count  >  total_face / 2)
  {
     response  =  "The seminar was engaging";
  }
  else
  {
     response  =  "The seminar was not engaging";
  }

Scenario 3: Populate Zoho Creator field with predicted emotion

Note: This scenario is only applicable for Zoho Creator.

Let’s consider a situation where you need to auto-populate predicted emotion on a single line field based on the image uploaded in the file upload / image field. The emotion can be fetched from the response of zoho.ai.detectFace task using get() function. This action can be performed with the help of Zoho Creator - Download File API.

The following script detects the emotion from the uploaded image and updates another Zoho Creator field with the predicted emotion:

// url - Contains the download file API I URL to download the uploaded image 
// account_owner_name - Replace with the  user name of the Creator account. 
// app_link_name - Replace with the link name of the target application. 
// report_link_name - Replace with the link name of the target report. 
// record_ID - Replace with the ID of the record from which you want to download the file. 
// field_link_name - Replace with the link name of the target file upload or image field. 

 image_file  =  invokeurl 
     [ 
       url: "https://creator.zoho.com/api/v2/<account_owner_name>/<app_link_name>/report/<report_link_name>/<record_ID>/<field_link_name>/download" 
       type : GET 
       connection: "zoho_oauth_connection" 
      ]; 
       
// Create a variable to store the data of prominent face 
detect_emotion  =  zoho.ai.detectFace(image_file,true,{"emotion"}).getJson("data").getJson("emotion").getJson("prediction"); 

// Specify a form field where the emotion needs to be populated. 
// Replace "predicted_emotion" with your single line field_link_name . 
 input.predicted_emotion  =  detect_emotion; 

Response Format

Success Response

  • The success response will be returned in the following format.

    {    "data": [
         {
          "faceCoordinate": [
            "0.565",
            "0.273",
            "0.760",
            "0.590"
          ],
          "gender": {
            "prediction": "male",
            "confidence": {
              "female": "0.028",
              "male": "0.972"
            }
          },
          "emotion": {
            "prediction": "happy",
            "confidence": {
              "angry": "0.0",
              "fear": "0.0",
              "happy": "0.999",
              "neutral": "0.0",
              "sad": "0.0",
              "surprise": "0.0"
            }
          },
          "pose": {
            "yaw": "-4.6450",
            "pitch": "-7.8978",
            "roll": "-1.0229"
          }
        }
      ],
      "numFacesFound": 1,
      "executionMessage": "Faces found",
      "message": "success"
    }

 

Failure Response

  • The error message when the file size exceeded the allowed limit will be returned in the following format:

     
     "The zoho.ai.detectFace integration task failed because the size of the supplied input file exceeded the allowed limit of 5MB."
     
  • The failure response when the image is not in the specified format(.png, .jpg,or .jpeg).

    {    "error":{
        "code": 400,
        "message": "null file format is not allowed"
      },
      "status": "Failed"
    }
     

 

Related Links

Get Started Now

Execute