PAEFF: Precise Alignment and Enhanced Gated Feature Fusion for Face-Voice Association
This work addresses a specific multimodal learning task for the AI community, with incremental improvements in face-voice association.
The paper tackles the problem of learning associations between faces and voices by addressing issues with negative mining and margin parameters, proposing a method that aligns embedding spaces and uses enhanced gated fusion, achieving improved performance on the VoxCeleb dataset.
We study the task of learning association between faces and voices, which is gaining interest in the multimodal community lately. These methods suffer from the deliberate crafting of negative mining procedures as well as the reliance on the distant margin parameter. These issues are addressed by learning a joint embedding space in which orthogonality constraints are applied to the fused embeddings of faces and voices. However, embedding spaces of faces and voices possess different characteristics and require spaces to be aligned before fusing them. To this end, we propose a method that accurately aligns the embedding spaces and fuses them with an enhanced gated fusion thereby improving the performance of face-voice association. Extensive experiments on the VoxCeleb dataset reveals the merits of the proposed approach.