AI Paper Atlas

Research map / Video / Video generation

The devil is in the details: Enhancing Video Virtual Try-On via Keyframe-Driven Details Injection

Qingdong He, Xueqin Chen, Yanjie Pan, Peng Tang, Pengcheng Xu, Zhenye Gan, Chengjie Wang, Xiaobin Hu, Jiangning Zhang, Yabiao Wang

CVPR 2026 · Video › Video generation · approach: Efficient video diffusion models · 2 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
add garment distillation component D (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 22.5241 -> 7.5267Improved
add garment details enhancement module (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 19.8968 -> 7.5267Improved
add collaborative background optimization (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 17.2134 -> 7.5267Improved
add instruction-guided keyframe sampling (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 16.2586 -> 7.5267Improved
add garment-area ratio score Sr(f) (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 17.2568 -> 7.5267Improved
use multiple keyframes instead of one (ablation)KeyTailor · ViT-HD ablationVFID_I (paired): 16.3856 -> 7.5267Improved
KeyTailor vs MagicTryOnKeyTailor · ViT-HD testVFID_I (paired): 14.0587 -> 7.5267Improved
fine-tune additionally on ViT-HDKeyTailor · ViViDVFID_I (paired): 9.5285 -> 8.2164Improved
fine-tune DreamVVT on own data vs KeyTailorKeyTailor vs DreamVVT · ViViDVFID_R (paired): 0.2549 vs 0.2761Mixed
use LoRA detail injection instead of added interaction modulesKeyTailor · Trainable paramsparams (B): 16.4460 -> 0.2057Improved
KeyTailor on image try-onKeyTailor · VITON-HDLPIPS: 0.0477 -> 0.0566Hurt

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