2025/03/11 by Jiaoqun Li, S. Kevin Zhou, Li, Junzhe +17 · 2 citations
Computer Science · #Face recognition and analysis #Face and Expression Recognition
paper · pdf · doi:10.48550/arxiv.2503.08120
Unified multimodal models (UMMs) have emerged as a powerful paradigm in fundamental cross-modality research, demonstrating significant potential in both image understanding and generation. However, existing research in the face domain primarily faces two challenges: (1) fragmentation development, with existing methods failing to unify understanding and generation into a single one, hindering the way to artificial general intelligence. (2) lack of fine-grained facial attributes, which are crucial for high-fidelity applications. To handle those issues, we propose UniF2ace, the first UMM specifically tailored for fine-grained face understanding and generation. First, we introduce a novel theoretical framework with a Dual Discrete Diffusion (D3Diff) loss, unifying masked generative models with discrete score matching diffusion and leading to a more precise approximation of the negative log-likelihood. Moreover, this D3Diff significantly enhances the model's ability to synthesize high-fidelity facial details aligned with text input. Second, we propose a multi-level grouped Mixture-of-Experts architecture, adaptively incorporating the semantic and identity facial embeddings to complement the attribute forgotten phenomenon in representation evolvement. Finally, to this end, we construct UniF2aceD-1M, a large-scale dataset comprising 130K fine-grained image-caption pairs and 1M visual question-answering pairs, spanning a much wider range of facial attributes than existing datasets. Extensive experiments demonstrate that UniF2ace outperforms existing models with a similar scale in both understanding and generation tasks, with 7.1% higher Desc-GPT and 6.6% higher VQA-score, respectively.