AI Paper Atlas

Research map / Image generation / Editing and style transfer

Wear-Any-Way: Manipulable Virtual Try-on via Sparse Correspondence Alignment

Mengting Chen, Xi Chen, Zhonghua Zhai, Chen Ju, Xuewen Hong, Jinsong Lan, Shuai Xiao

ECCV 2024 · Image generation › Editing and style transfer · approach: Virtual try-on · 37 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
inject point embedding into attention q,k instead of latent noise (ablation)Wear-Any-Way · Point control benchmark (upper)landmark distance: 27.32 -> 24.35Improved
add point embedding at all (attention injection) (ablation)Wear-Any-Way · Point control benchmark (upper)landmark distance: 35.65 -> 24.35Improved
add zero-initialization to point embedding output (ablation)Wear-Any-Way · Point control benchmark (upper)landmark distance: 24.35 -> 22.65Improved
add condition dropping (drop pose map, box-ify mask) (ablation)Wear-Any-Way · Point control benchmark (upper)landmark distance: 22.65 -> 18.39Improved
add point-weighted loss (ablation)Wear-Any-Way · Point control benchmark (upper)landmark distance: 18.39 -> 17.65Improved
use Stable Diffusion features for training point matching instead of CLIP or DINO (ablation)point-pair collection · Training point-pair matching (upper)landmark distance: 93.42 (CLIP), 83.24 (DINO) -> 43.44Improved
use reference U-Net rather than CLIP/DINOv2/ControlNet to encode the garment (ablation)Wear-Any-Way try-on baseline · VITON-HD (qualitative)detail preservationImproved
Wear-Any-Way vs StableVITON on DressCode upperWear-Any-Way · DressCode upper-bodyFID: 11.72 vs 11.266 (FID), 0.33 vs 0.72 (KID)Mixed

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 0 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy