VersaVogue: Visual Expert Orchestration and Preference Alignment for Unified Fashion Synthesis

TL;DR

VersaVogue introduces a unified diffusion-based framework for fashion image synthesis that enables disentangled, multi-condition control and improves realism without relying on human annotations.

Contribution

It proposes a trait-routing attention module for dynamic feature injection and an automated preference optimization pipeline for enhanced controllability and realism.

Findings

Outperforms existing methods in visual fidelity and semantic consistency.

Supports both garment generation and virtual dressing in a unified framework.

Achieves fine-grained attribute control without human-labeled data.

Abstract

Diffusion models have driven remarkable advancements in fashion image generation, yet prior works usually treat garment generation and virtual dressing as separate problems, limiting their flexibility in real-world fashion workflows. Moreover, fashion image synthesis under multi-source heterogeneous conditions remains challenging, as existing methods typically rely on simple feature concatenation or static layer-wise injection, which often causes attribute entanglement and semantic interference. To address these issues, we propose VersaVogue, a unified framework for multi-condition controllable fashion synthesis that jointly supports garment generation and virtual dressing, corresponding to the design and showcase stages of the fashion lifecycle. Specifically, we introduce a trait-routing attention (TA) module that leverages a mixture-of-experts mechanism to dynamically route condition features to the most compatible experts and generative layers, enabling disentangled injection of visual attributes such as texture, shape, and color. To further improve realism and controllability, we develop an automated multi-perspective preference optimization (MPO) pipeline that constructs preference data without human annotation or task-specific reward models. By combining evaluators of content fidelity, textual alignment, and perceptual quality, MPO identifies reliable preference pairs, which are then used to optimize the model via direct preference optimization (DPO). Extensive experiments on both garment generation and virtual dressing benchmarks demonstrate that VersaVogue consistently outperforms existing methods in visual fidelity, semantic consistency, and fine-grained controllability.