Face recognition in video content
First Claim
1. In a computing environment, a method performed at least in part on at least one processor, comprising, receiving face detection data corresponding to a face detected in an input video frame, building face galleries, including grouping faces detected in input video frames into candidate groups based upon similarity data, filtering at least some faces from a candidate group, adding remaining faces to one of the face galleries, and labeling each face gallery with the face identification data, matching the face detection data against face identification data maintained in a face gallery among a plurality of face galleries to recognize the face in the input video frame, and generating metadata that associates the video frame and the face with the face identification data.
4 Assignments
0 Petitions
Accused Products
Abstract
The subject disclosure relates to face recognition in video. Face detection data in frames of input data are used to generate face galleries, which are labeled and used in recognizing faces throughout the video. Metadata that associates the video frame and the face are generated and maintained for subsequent identification. Faces other than those found by face detection may be found by face tracking, in which facial landmarks found by the face detection are used to track a face over previous and/or subsequent video frames. Once generated, the maintained metadata may be accessed to efficiently determine the identity of a person corresponding to a viewer-selected face.
-
Citations
17 Claims
- 1. In a computing environment, a method performed at least in part on at least one processor, comprising, receiving face detection data corresponding to a face detected in an input video frame, building face galleries, including grouping faces detected in input video frames into candidate groups based upon similarity data, filtering at least some faces from a candidate group, adding remaining faces to one of the face galleries, and labeling each face gallery with the face identification data, matching the face detection data against face identification data maintained in a face gallery among a plurality of face galleries to recognize the face in the input video frame, and generating metadata that associates the video frame and the face with the face identification data.
- 8. In a computing environment, a system comprising, a face recognition pipeline that recognizes faces from input video, including a face grouping module configured to group faces into groups by similarity based upon face detection data provided by a face detection module data, the grouping module further configured to provide face galleries corresponding to the groups, including information that identifies each person associated with a face in a face gallery, the face recognition pipeline further comprising a face recognition mechanism that matches faces in the input video with faces in the face galleries to output information corresponding to recognized faces in the input video, wherein the input video comprises a full set of episodes, wherein the face grouping module groups faces using a lesser subset of the episodes, and wherein the face recognition mechanism matches faces for the full set of episodes.
-
15. One or more computer-readable storage media having computer-executable instructions, which when executed perform steps, comprising:
-
receiving face detection data corresponding to a face detected in an input video frame, wherein the face detection data corresponds to similarity data; tracking the face in one or more adjacent video frames based on at least some of the face detection data to acquire a tracked face; and utilizing the similarity data to determine whether the tracked face matches a threshold level of similarity; and maintaining the tracked face in a face gallery in a single candidate group, wherein the tracked face from each video frame among the one or more adjacent video frames is maintained in the face gallery. - View Dependent Claims (16, 17)
-
Specification