System, method, and recording medium for detecting video face clustering with inherent and weak supervision

A face clustering method for video face clustering in a video sequence, the method comprising:

• collecting group-level supervision and instance level supervision within a same chunklet based on a user input of face images for a person;
• and
• in an embedding space, executing pair-wise based clustering to cluster the video images into different clusters with the instance level supervision collected by the collecting by discriminatively learning linear or nonlinear transformations between the face images, and
• wherein the executing further performs integration of pairwise must-link and must-not-link constraints into a process of model fitting and penalizes a constraint violation whenever latent variables in a constraint are different while they are supposed to be the same.

A non-transitory computer-readable recording medium recording a face clustering detection program for video face clustering in a video sequence, the program causing a computer to perform:

• collecting group-level supervision and instance level supervision within a same chunklet based on a user input of face images for a person;
• and
• in an embedding space, executing pair-wise based clustering to cluster the video images into different clusters with the instance level supervision collected by the inherent supervision summarization device by discriminatively learning linear or nonlinear transformations between the face images, and
• wherein the executing further performs integration of pairwise must-link and must-not-link constraints into a process of model fitting and penalizes a constraint violation whenever latent variables in a constraint are different while they are supposed to be the same.