Paper 2601.01425
DreamID-V:Bridging the Image-to-Video Gap for High-Fidelity Face Swapping via Diffusion Transformer
- Published
- Jan 2026
- Research lab
- Independent
- Citations
- 7
- GitHub
- 668 stars
01 In brief
Summary
The paper introduces DreamID-V, a Diffusion Transformer (DiT)-based framework for high-fidelity video face swapping (VFS).
It addresses the gap between image face swapping (IFS) and VFS by proposing a data pipeline, SyncID-Pipe, which pre-trains an Identity-Anchored Video Synthesizer (IVS) to generate synthetic videos, combined with IFS models to create bidirectional ID quadruplets for explicit supervision.
The framework uses a Modality-Aware Conditioning mechanism to inject spatio-temporal context, structural guidance, and identity information.
Training includes a Synthetic-to-Real Curriculum and Identity-Coherence Reinforcement Learning (IRL) to enhance realism and identity consistency.
The authors also introduce IDBench-V, a benchmark with 200 real-world pairs.
Experiments show DreamID-V outperforms state-of-the-art methods in identity similarity, attribute preservation, and video quality, and can be adapted to other swap tasks like outfit or hairstyle swapping.
02 From the paper
Abstract
Video Face Swapping (VFS) requires seamlessly injecting a source identity into a target video while meticulously preserving the original pose, expression, lighting, background, and dynamic information. Existing methods struggle to maintain identity similarity and attribute preservation while preserving temporal consistency. To address the challenge, we propose a comprehensive framework to seamlessly transfer the superiority of Image Face Swapping (IFS) to the video domain. We first introduce a novel data pipeline SyncID-Pipe that pre-trains an Identity-Anchored Video Synthesizer and combines it with IFS models to construct bidirectional ID quadruplets for explicit supervision. Building upon paired data, we propose the first Diffusion Transformer-based framework DreamID-V, employing a core Modality-Aware Conditioning module to discriminatively inject multi-model conditions. Meanwhile, we propose a Synthetic-to-Real Curriculum mechanism and an Identity-Coherence Reinforcement Learning strategy to enhance visual realism and identity consistency under challenging scenarios. To address the issue of limited benchmarks, we introduce IDBench-V, a comprehensive benchmark encompassing diverse scenes. Extensive experiments demonstrate DreamID-V outperforms state-of-the-art methods and further exhibits exceptional versatility, which can be seamlessly adapted to various swap-related tasks.