CORAL: Correspondence Alignment for Improved Virtual Try-On
2026-02-19 • Computer Vision and Pattern Recognition
Computer Vision and Pattern Recognition
AI summaryⓘ
The authors study how current virtual try-on methods struggle to keep clothing details because they don't properly match the person and garment features. They find that in diffusion transformer models, good matching depends on correctly pairing queries and keys during attention. To fix this, the authors create CORAL, a system that improves alignment by using a special loss to guide matches and another to make the attention more focused. Their approach improves how well the clothing fits and looks in virtual try-on tests, as shown in their experiments.
Virtual Try-OnDiffusion TransformersAttention MechanismQuery-Key MatchingPerson-Garment CorrespondenceLoss FunctionsEntropy MinimizationCorrespondence Distillation3D AttentionVision-Language Models
Authors
Jiyoung Kim, Youngjin Shin, Siyoon Jin, Dahyun Chung, Jisu Nam, Tongmin Kim, Jongjae Park, Hyeonwoo Kang, Seungryong Kim
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.