The face clustering feature groups images in a media set that contain similar faces. This feature is useful for scenarios such as creating face-based photo albums in cloud storage, detecting strangers in home monitoring systems, and managing customers in new retail. After the faces are clustered, you can query all images of a person based on their face group.
This document is no longer maintained. Use the new version of Intelligent Media Management.
For a comparison of the new and legacy versions of Intelligent Media Management, see Version usage guide.
For more information about the face clustering feature in the new version of Intelligent Media Management, see Face clustering.
To enable face clustering, use an SDK in your preferred language to call the following API operations in sequence.
Call the PutProject API operation to create a project.
If a project already exists, skip this step. You can call the ListProjects API operation to list all projects in a specified region.
NoteAlternatively, you can create a project in the Intelligent Media Management console. For more information, see Create a project.
Call the CreateSet API operation to create a media set and obtain the media set ID (SetId).
If a suitable media set already exists, skip this step. You can call the ListSets API operation to obtain information about all media sets in the specified project.
Call the IndexImage API operation to index images into the media set.
If you have many images, repeat this operation as needed.
Call the CreateGroupFacesJob API operation to cluster the faces in the media set.
You can repeat Step 3 and Step 4 to process incremental face clustering.
Query all images of a person based on their face group.
Call the ListFaceGroups API operation to obtain information about the face groups in the media set.
Call the FindImages API operation to query all images of the person that correspond to the face group.