AI Paper Atlas

Research map / Video / Video generation

Vivid-VR: Distilling Concepts from Text-to-Video Diffusion Transformer for Photorealistic Video Restoration

Haoran Bai, Xiaoxu Chen, Canqian Yang, Zongyao He, Sibin Deng, Ying Chen

ICLR 2026 · Video › Video generation · approach: Efficient video diffusion models

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 concept distillation (100K T2V-generated training videos) (ablation)Vivid-VR · UGC50MUSIQ: 57.36 -> 67.61Improved
add concept distillation (ablation)Vivid-VR · UGC50NIQE: 5.364 -> 4.361Improved
use Qwen2.5-VL captioner instead of concept distillation (ablation)Vivid-VR · UGC50MUSIQ: 67.61 -> 60.88Hurt
add control feature projector (ablation)Vivid-VR · UGC50MUSIQ: 63.06 -> 67.61Improved
fine-tune VAE encoder instead of projector (ablation)Vivid-VR · UGC50MUSIQ: 67.61 -> 64.31Hurt
add cross-attention branch to MLP connector (ablation)Vivid-VR · UGC50MUSIQ: 59.78 -> 67.61Improved
keep MLP branch instead of skip connection (ablation)Vivid-VR · UGC50MUSIQ: 63.91 -> 67.61Improved
use dual-branch connector instead of ZeroSFT (ablation)Vivid-VR · UGC50MUSIQ: 61.21 -> 67.61Improved
remove MLP branch entirely (ablation)Vivid-VR · UGC50convergenceHurt
increase generated training videos from 0 to 100K (ablation)Vivid-VR · UGC50MUSIQ: 57.36 -> 67.61Improved
increase generated training videos from 100K to 150K (ablation)Vivid-VR · UGC50MUSIQ: 67.61 -> 67.19No clear effect
train on generated videos only (no source videos) (ablation)Vivid-VR · UGC50MUSIQ: 67.19 -> 53.77Hurt

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