Research map / Image generation / Text-to-image diffusion
MMFace-DiT: A Dual-Stream Diffusion Transformer for High-Fidelity Multimodal Face Generation
Bharath Krishnamurthy, Ajita Rattani
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add Modality Embedder to the baseline DiT (ablation) | baseline DiT (SD2 VAE) · Text + Mask face generation | FID: 44.52 -> 40.49 | Improved |
| add Dual-Stream design (ablation) | DiT + Modality Embedder · Text + Mask face generation | CLIP score: 24.53 -> 29.69 | Improved |
| add Dual-Stream design (ablation) | DiT + Modality Embedder · Text + Mask face generation | mIoU: 46.34 -> 48.91 | Improved |
| replace standard attention with shared RoPE attention (ablation) | DiT + ME + Dual-Stream · Text + Mask face generation | mIoU: 48.91 -> 50.05 | Improved |
| add Modality Embedder (no extra training per modality) (ablation) | baseline DiT (SD2 VAE) · Text + Mask face generation | CLIP score: 24.53 -> 24.31 | No clear effect |
| replace SD2 VAE with Flux VAE (ablation) | MMFace-DiT · Text + Mask face generation | FID: 33.77 -> 27.95 | Improved |
| replace SD2 VAE with Flux VAE (ablation) | MMFace-DiT · Text + Mask face generation | mIoU: 50.05 -> 49.16 | Hurt |
| use SD3 VAE instead of Flux VAE (ablation) | MMFace-DiT · Text + Mask face generation | FID: 27.95 -> 20.73 | Improved |
| use SD3 VAE instead of Flux VAE (ablation) | MMFace-DiT · Text + Mask face generation | LPIPS: 0.340 -> 0.368 | Hurt |
| use Qwen-Image VAE instead of SDXL/SD2 (ablation) | MMFace-DiT · Text + Mask face generation | FID: 23.00 -> 33.09 | Hurt |
| train with rectified flow matching instead of DDPM (ablation) | MMFace-DiT · Text + Sketch face generation | FID: 27.67 -> 9.14 | Improved |
| train with rectified flow matching instead of DDPM (ablation) | MMFace-DiT · Text + Mask face generation | FID: 27.95 -> 16.63 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 2 of 2 on most shared results.
- ControlNet — ControlNet++: Improving Conditional Controls with Efficient Consistency Feedback (ECCV 2024) · better on 4 of 4 shared results
- TediGAN — TediGAN: Text-Guided Diverse Face Image Generation and Manipulation (CVPR 2021) · better on 2 of 2 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- TrioPose: Native Triple-Stream Diffusion Transformers for Pose-Guided Text-to-Image Generation (NeurIPS 2026 · similarity 0.89)
- Text-Guided 3D Face Synthesis - From Generation to Editing (CVPR 2024 · similarity 0.89)
- Exploring the Deep Fusion of Large Language Models and Diffusion Transformers for Text-to-Image Synthesis (CVPR 2025 · similarity 0.88)
- Fast Text-to-3D-Aware Face Generation and Manipulation via Direct Cross-modal Mapping and Geometric Regularization (ICML 2024 · similarity 0.88)
- DuoGen: Towards Autonomous Interleaved Multimodal Generation (CVPR 2026 · similarity 0.87)
- When StyleGAN Meets Stable Diffusion: a W+ Adapter for Personalized Image Generation (CVPR 2024 · similarity 0.87)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
