How to set up Image Proctoring?

Know to set up Image Proctoring to capture real-time images of the candidates when they appear for the test

iMocha uses advanced AI-enabled cheating prevention measures to ensure that the candidates take the tests fairly. Image Proctoring is an online invigilation that helps you get the candidate's images when the Test is being attempted. Online proctoring of the candidate is conducted remotely by the system in real time. Live images of the candidate are taken at regular intervals through their system's webcam. The system reports if the candidate has used unfair means to take the Test.

To enable Image Proctoring at Global Level:

1. Click Settings -> Global Settings
2. Click Master Test Settings from the list on the left
3. Click Proctoring Settings
   
4. Select Image Proctoring
5. You can choose to terminate the test along with the maximum number of image violations, as the test will be terminated if the candidate exceeds the maximum number of image violations.
6. Click Save

To enable Image Proctoring at Test Level:

1. Click My Tests 
2. Select the Test for which you want to enable Image Proctoring
3. Click Settings > Proctoring Settings
   
   
4. You can choose to terminate the test along with the maximum number of image violations, as the test will be terminated if the candidate exceeds the maximum number of image violations.
5. Click Save

Ensure to enable Image Proctoring when you send email invites/open links to candidates

Image Proctoring will be enabled for this test, and you can see the images in Reports under Proctoring Details tab. 

How does Image Proctoring work?

The system captures a base profile image when the candidate starts the test. 

Using this image as a reference, the system will display the following image violation conditions:

• Verified Face

The image matches the base profile image. No violation.

• No Face

The camera could not detect the candidate's face, or the candidate's face was covered.

• Multiple Face

Two or more faces were detected along with the base image.

• Unrecognized Face

Someone other than the candidate appears to be taking the test. The system compares the image against the base profile image to flag this.

Reviewing the context around a flagged violation

When an image is flagged as No Face, Multiple Face, or Unrecognized Face, you no longer have to judge it from that one image alone. iMocha now captures the images taken just before and after the flagged moment, so you can see what actually happened.

• In the Proctoring Details grid, a flagged thumbnail shows a counter badge (for example, "4/7") if surrounding images are available. Hover over the thumbnail to preview those surrounding images.
  
  
• Click the thumbnail to open the full view. It opens on the flagged image by default, with a filmstrip below it so you can step through every image captured around that moment.
  
  
• If a test has more than one flagged moment, you can browse between them from the same view without closing and reopening it.

This applies to tests completed after this update. For tests completed earlier, flagged images continue to display as a single image, exactly as before.

Understanding the Proctoring Test Report

The report contains the following:

• Color Coded Warning Flags to identify low, medium, and high cheating intensity at first glance.
• Color Coded Warning icon to indicate the intensity of the number of Window Violations, along with the time duration for each violation.

iMocha’s Image Proctoring solution considers tolerance levels for different types of cheating and raises precise warning flags. The final cheating detection report that comes to your table has eliminated misinterpretations of fraud instances.

Following is the logic behind Color Coded Warning Flags that appear in the Proctoring Details within the Test Reports. 

You can use the above table to understand the specific color of the warning flag that are indicated in the proctoring report. 
The system already adjusts any scope of errors captured during cheating instances so that only a genuine count of violations is considered for decision-making. 

For any queries, mail us at support@imocha.io.