Stable-Hair v2: Real-World Hair Transfer via Multiple-View Diffusion Model

TL;DR

Stable-Hair v2 introduces a multi-view diffusion framework for high-fidelity, view-consistent hair transfer, leveraging novel data generation and training strategies to outperform existing methods in multi-view realism.

Contribution

It is the first to utilize multi-view diffusion models for robust, high-quality, view-consistent hair transfer across multiple perspectives.

Findings

Achieves seamless, realistic multi-view hair transfer

Outperforms existing methods in view consistency and detail

Establishes a new benchmark in multi-view hair transfer

Abstract

While diffusion-based methods have shown impressive capabilities in capturing diverse and complex hairstyles, their ability to generate consistent and high-quality multi-view outputs -- crucial for real-world applications such as digital humans and virtual avatars -- remains underexplored. In this paper, we propose Stable-Hair v2, a novel diffusion-based multi-view hair transfer framework. To the best of our knowledge, this is the first work to leverage multi-view diffusion models for robust, high-fidelity, and view-consistent hair transfer across multiple perspectives. We introduce a comprehensive multi-view training data generation pipeline comprising a diffusion-based Bald Converter, a data-augment inpainting model, and a face-finetuned multi-view diffusion model to generate high-quality triplet data, including bald images, reference hairstyles, and view-aligned source-bald pairs. Our multi-view hair transfer model integrates polar-azimuth embeddings for pose conditioning and temporal attention layers to ensure smooth transitions between views. To optimize this model, we design a novel multi-stage training strategy consisting of pose-controllable latent IdentityNet training, hair extractor training, and temporal attention training. Extensive experiments demonstrate that our method accurately transfers detailed and realistic hairstyles to source subjects while achieving seamless and consistent results across views, significantly outperforming existing methods and establishing a new benchmark in multi-view hair transfer. Code is publicly available at https://github.com/sunkymepro/StableHairV2.