AI Paper Atlas

Research map / Video / Video generation

VideoVAE+: Large Motion Video Autoencoding with Cross-modal Video VAE

Yazhou Xing, Yang Fei, Yingqing He, Jingye Chen, Jiaxin Xie, Xiaowei Chi, Qifeng Chen

ICCV 2025 · Video › Video generation · approach: Autoregressive long video generation · 4 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
use two-stage temporal-aware spatial plus temporal compression instead of simultaneous (ablation)Video VAE · Large Motion testPSNR: 24.0593 -> 24.6722Improved
use two-stage compression instead of sequential (ablation)Video VAE · Large Motion testPSNR: 23.3681 -> 24.6722Improved
use two-stage compression instead of simultaneous (ablation)Video VAE · Large Motion testLPIPS: 0.1293 -> 0.1162Improved
use two-stage compression instead of simultaneous (ablation)Video VAE · Large Motion testSSIM: 0.7315 -> 0.7234Hurt
add image-video joint training (ablation)Video VAE 4-channel · WebVid testPSNR: 30.2091 -> 30.3140Improved
add image-video joint training (ablation)Video VAE 16-channel · WebVid testPSNR: 33.8844 -> 34.1558Improved
add image-video joint training (ablation)Video VAE 4-channel · Image reconstructionPSNR: 15.1001 -> 30.8650Improved
add image-video joint training (ablation)Video VAE 16-channel · Image reconstructionPSNR: 9.2603 -> 35.3437Improved
add cross-modal text guidance (ablation)Video VAE 16-channel · WebVid testPSNR: 34.1558 -> 34.5022Improved
add cross-modal text guidance (ablation)Video VAE 4-channel · WebVid testPSNR: 30.3140 -> 30.1110Hurt
add cross-modal text guidance (ablation)Video VAE 16-channel · WebVid testLPIPS: 0.0271 -> 0.0323Hurt
use video GAN loss instead of image GAN loss (ablation)temporal-aware spatial autoencoder · MMTrailer valPSNR: 31.9133 -> 32.0262Improved

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