Facial recognition refers to a set of services provided by the Content Detection API based on face image detection technology. It includes facial attribute detection, face comparison, and face retrieval to meet your face image detection requirements. This topic describes the features and key concepts of facial recognition.
Features
Feature | Description | Use case | API operations |
Facial attribute detection | Detects one or more faces in an image and analyzes attributes such as expressions. | Typically used to detect whether there are faces in an image and how many. | |
Face comparison | Determines whether the faces in two given images belong to the same person. Currently, only images with a single face are supported. | Typically used for attendance check-in scenarios. | |
Face retrieval | Retrieves a list of individuals from a given face image collection whose faces are most similar to the face in the given image. | Typically used for identity verification at security gates. |
Key concepts
In facial recognition technology, you need to understand the following concepts:
Face: In facial recognition technology, this refers to a face detected from an image to be processed.
When the system performs facial recognition on an image, it records the detected faces, including their position information in the image. Typically, multiple faces in the same image can be detected as long as there are sufficient facial landmarks.
Person: In facial recognition technology, this represents the individual associated with a particular face image.
Person information is typically specified by the user when uploading a face image. Usually, all face images associated with the same person must belong to the same individual, and each image must contain no more than one face.
Group: In facial recognition technology, this is a logical group used to store person information.
When users need to retrieve faces from a given collection that are identical or similar to a given face image, they must specify a group.
A person can belong to one or more groups, and a person can have multiple images as their face library.