Detecting, Analyzing, and Recognizing Faces

In addition to detecting a face in an image with a bounding box, what other capabilities does the Face service offer over the Computer Vision service?

• Facial recognition

• Comprehensive facial features (e.g. emotional state)

• Age

• Gender

Using the Computer Vision service, how can you get information (e.g. age and gender) for more than one person in a single image?

• You must send the image through multiple times, blocking out all but one face for each submission.

• The Computer Vision service will return a JSON object that contains an array of faces.

• The Computer Vision service will return a JSON object that contains a list of face locations, and you must then specify which location you want additional information on.

• You can only get information on one person per image.

Can you use Face services to verify that two images are of the same person anonymously?

• Yes, the Face service assigns a unique ID to each face. This can be used to verify without needing additional identifiable information.

• Partially, while the Face service can anonymously detect that two faces are similar, it cannot verify that they are the same without a human verifying additional personally identifiable information.

• . No, while the Face service is able to verify that two faces are the same, it can only do this by also storing additional identifiable information (e.g. name)

• No, the Face service isn’t able to verify that two faces are the same.

Can Face services be used to identify a criminal based on a sketch made by a sketch artist?

• No, even if it were possible to have an accurate comparison, it is not permitted within the EULA.

• Yes, with a combination of Face services and Custom Vision a face-to-sketch comparison is possible.

• Yes, and this is commonly used.

• No, while it is possible to compare a face with a sketch, it is not permitted within the EULA

“Facial landmark location” is a feature of Face services that identifies the location in the image of what?

• The boundaries of the face (e.g. the top, sides, and bottom of each face in the image).

• Faces that are on popular landmarks (e.g. the Statue of Liberty).

• The specific facial features of a face (e.g. the tip of the nose).

• The landmarks that are near to the face (e.g. if a person is standing in front of the Eiffel Tower).

Describe one consideration for the ethical use of face analysis in applications.

Answer explanation

Your Answer: Should include either Data privacy and security, Transparency, or Fairness and inclusiveness. For Data privacy and security, might mention that facial data is personally identifiable and should be considered sensitive and private. For Transparency, might mention that users need to be informed of how their facial data will be used and who will have access to it. For fairness and inclusiveness, might include that a face-based system cannot be user in a manner that is prejudicial to individuals based on their appearance, or to unfairly target individuals.

A theme park takes 20 images of each rider while they ride down a roller coaster. Rather than display all 20 to the customers, they decide to use Face service to show the best 5. What functionality should they focus on for choosing the top 5 images?

Answer explanation

Your Answer: Should likely mention the Face Attribute Analysis, specifically head pose, blur, noise, and occlusion. Might also mention emotions and smile.

LusuTechnik wants to train a facial recognition model on their authorized employees for a particular lab. Describe the steps for creating this application using the Face services.

Answer explanation

Your Answer: Should mention creating a Person Group for the authorized employees, creating a Person within that Person Group for each individual you want represented, and then adding detected faces from multiple images of that individual to the Person. After this, they should train and deploy the model.

Amina Binti Amani Hospital wants to verify that the guest who checked in at reception is the same guest who entered a particular room. What should you do to solve this challenge?

Answer explanation

Your Answer: Should mention verifying the face detected at the room with the face ID of the face detected at reception.

You need to create a facial recognition solution to identify specific employees at a company. Should you use Custom Vision or Face services for your solution? Why?

Answer explanation

Your Answer: Should include Face because with Face services you can create a facial recognition solution.