AI Paper Atlas

Research map / Video / Video generation

OpenVid++: A Large-scale High-quality Dataset for Text-to-video Generation

Kepan Nan, Tiehan Fan, Rui Xie, Penghao Zhou, Yilin Zhang, Xiang Li, Jian Yang, Zhenheng Yang, Ying Yu Tai

IJCV 2026 · Video › Video generation · approach: Compositional text-to-video diffusion · 7 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
replace STDiT with MVDiT (ablation)T2V model, 256x256, OpenVid-0.4M · T2V benchmarkVQA_A: 11.11 -> 22.39Improved
increase resolution 256 to 512 for STDiT (ablation)STDiT · T2V benchmarkVQA_A: 11.11 -> 65.15Improved
train MVDiT on OpenVid-1M instead of OpenVid-0.4M (ablation)MVDiT-256 · T2V benchmarkVQA_A: 22.39 -> 24.87Improved
add multi-head cross-attention (MHCA) (ablation)MVDiT-256 · T2V benchmarkVQA_A: 13.9 -> 22.39Improved
add scaling parameter alpha (ablation)MVDiT-256 · T2V benchmarkVQA_A: 3.16 -> 22.39Improved
add clarity filtering to aesthetics, temporal and motion filters (ablation)data pipeline · T2V benchmark (0.6M, 256x256)VQA_A: 20.32 -> 30.26Improved
add clarity filtering to other three filters (ablation)data pipeline · T2V benchmark (0.6M, 256x256)Blip_bleu: 24.43 -> 23.43Hurt
train on OpenVid-1M instead of WebVid-10MSTDiT 256x256 · T2V benchmarkVQA_A: 13.40 -> 17.78Improved
train on OpenVid-1M instead of Panda-50MSTDiT 256x256 · T2V benchmarkVQA_T: 9.60 -> 12.98Improved
train on OpenVidHD-0.4M instead of Panda-50M-HDSTDiT 1024x1024 · T2V benchmarkVQA_A: 13.48 -> 73.46Improved
train on OpenVidHD-0.4M instead of WebVid-10M with super-resolutionSTDiT 1024x1024 · T2V benchmarkWarping_error: 0.0137 -> 0.0052Improved
train restoration model on OpenVidHD-0.4M data instead of WebVid/YouTube dataI2VGen-XL video restoration · UDM10LPIPS: 0.2731 -> 0.2015Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 2 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