LEED: Label-Free Expression Editing via Disentanglement
This work addresses the challenge of reducing reliance on labeled data for facial expression editing, which is incremental by building on existing disentanglement techniques to enable label-free operation.
The paper tackles the problem of facial expression editing without requiring expensive expression labels by proposing a label-free framework that disentangles identity and expression attributes. The method achieves superior editing performance on both frontal and profile facial images, as demonstrated by extensive experiments on two public datasets.
Recent studies on facial expression editing have obtained very promising progress. On the other hand, existing methods face the constraint of requiring a large amount of expression labels which are often expensive and time-consuming to collect. This paper presents an innovative label-free expression editing via disentanglement (LEED) framework that is capable of editing the expression of both frontal and profile facial images without requiring any expression label. The idea is to disentangle the identity and expression of a facial image in the expression manifold, where the neutral face captures the identity attribute and the displacement between the neutral image and the expressive image captures the expression attribute. Two novel losses are designed for optimal expression disentanglement and consistent synthesis, including a mutual expression information loss that aims to extract pure expression-related features and a siamese loss that aims to enhance the expression similarity between the synthesized image and the reference image. Extensive experiments over two public facial expression datasets show that LEED achieves superior facial expression editing qualitatively and quantitatively.