CORAL: Correspondence Alignment for Improved Virtual Try-On

TL;DR

This paper introduces CORAL, a framework that explicitly aligns person-garment correspondence in virtual try-on using Diffusion Transformers, improving detail preservation and shape transfer in unpaired settings.

Contribution

The paper proposes CORAL, a novel DiT-based method with correspondence alignment and new evaluation protocol for better virtual try-on performance.

Findings

Improved garment detail preservation in VTON.

Enhanced global shape transfer accuracy.

Validated effectiveness through extensive ablations.

Abstract

Existing methods for Virtual Try-On (VTON) often struggle to preserve fine garment details, especially in unpaired settings where accurate person-garment correspondence is required. These methods do not explicitly enforce person-garment alignment and fail to explain how correspondence emerges within Diffusion Transformers (DiTs). In this paper, we first analyze full 3D attention in DiT-based architecture and reveal that the person-garment correspondence critically depends on precise person-garment query-key matching within the full 3D attention. Building on this insight, we then introduce CORrespondence ALignment (CORAL), a DiT-based framework that explicitly aligns query-key matching with robust external correspondences. CORAL integrates two complementary components: a correspondence distillation loss that aligns reliable matches with person-garment attention, and an entropy minimization loss that sharpens the attention distribution. We further propose a VLM-based evaluation protocol to better reflect human preference. CORAL consistently improves over the baseline, enhancing both global shape transfer and local detail preservation. Extensive ablations validate our design choices.