PerFace: Metric Learning in Perceptual Facial Similarity for Enhanced Face Anonymization
This work addresses privacy concerns in face anonymization by enabling better identity selection for face-swapping, though it is incremental as it builds on existing metric learning approaches.
The paper tackled the problem of measuring nuanced facial similarities for face anonymization by proposing a human-perception-based metric, resulting in significant improvements in similarity prediction and attribute-based classification tasks.
In response to rising societal awareness of privacy concerns, face anonymization techniques have advanced, including the emergence of face-swapping methods that replace one identity with another. Achieving a balance between anonymity and naturalness in face swapping requires careful selection of identities: overly similar faces compromise anonymity, while dissimilar ones reduce naturalness. Existing models, however, focus on binary identity classification "the same person or not", making it difficult to measure nuanced similarities such as "completely different" versus "highly similar but different." This paper proposes a human-perception-based face similarity metric, creating a dataset of 6,400 triplet annotations and metric learning to predict the similarity. Experimental results demonstrate significant improvements in both face similarity prediction and attribute-based face classification tasks over existing methods.