Cross-view Masked Diffusion Transformers for Person Image Synthesis

TL;DR

This paper introduces X-MDPT, a novel diffusion transformer model for pose-guided human image synthesis that outperforms existing methods in quality, efficiency, and parameter count by employing masked diffusion transformers on latent patches.

Contribution

The paper proposes a new diffusion model with masked transformers and a mask cross-prediction module, significantly improving synthesis quality and efficiency over prior Unet-based approaches.

Findings

Outperforms state-of-the-art on DeepFashion dataset

Achieves higher FID, SSIM, and LPIPS scores with fewer parameters

Faster inference and training efficiency

Abstract

We present X-MDPT ($\underline{Cross}$-view $\underline{M}$asked $\underline{D}$iffusion $\underline{P}$rediction $\underline{T}$ransformers), a novel diffusion model designed for pose-guided human image generation. X-MDPT distinguishes itself by employing masked diffusion transformers that operate on latent patches, a departure from the commonly-used Unet structures in existing works. The model comprises three key modules: 1) a denoising diffusion Transformer, 2) an aggregation network that consolidates conditions into a single vector for the diffusion process, and 3) a mask cross-prediction module that enhances representation learning with semantic information from the reference image. X-MDPT demonstrates scalability, improving FID, SSIM, and LPIPS with larger models. Despite its simple design, our model outperforms state-of-the-art approaches on the DeepFashion dataset while exhibiting efficiency in terms of training parameters, training time, and inference speed. Our compact 33MB model achieves an FID of 7.42, surpassing a prior Unet latent diffusion approach (FID 8.07) using only $11\times$ fewer parameters. Our best model surpasses the pixel-based diffusion with $\frac{2}{3}$ of the parameters and achieves $5.43 \times$ faster inference. The code is available at https://github.com/trungpx/xmdpt.