AI Paper Atlas

Research map / Video / Video generation

ConsID-Gen: View-Consistent and Identity-Preserving Image-to-Video Generation

Mingyang Wu, Ashirbad Mishra, Soumik Dey, Shuo Xing, Naveen Ravipati, Hansi Wu, Binbin Li, Zhengzhong Tu

CVPR 2026 · Video › Video generation · approach: Compositional text-to-video diffusion · 10 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
ConsID-Gen vs Wan2.2-5B on proprietary subsetI2V generation · ConsIDVid-Bench proprietary subsetI2V Subject consistency: 96.85 -> 98.31Improved
ConsID-Gen vs Wan2.1-14B on proprietary subsetI2V generation · ConsIDVid-Bench proprietary subsetMEt3R: 0.1572 -> 0.0978Improved
ConsID-Gen vs Wan2.1-14B on proprietary subsetI2V generation · ConsIDVid-Bench proprietary subsetChamfer Distance: 0.0866 -> 0.0996Hurt
ConsID-Gen vs Wan2.1-14B on public subsetI2V generation · ConsIDVid-Bench public subsetMEt3R: 0.0961 -> 0.1321Hurt
ConsID-Gen vs Wan2.2-5B on public subsetI2V generation · ConsIDVid-Bench public subsetI2V Subject consistency: 98.47 -> 98.14Hurt
ConsID-Gen vs Wan2.1-14B on public subsetI2V generation · ConsIDVid-Bench public subsetChamfer Distance: 0.1322 -> 0.1277Improved
add geometry encoder (VGGT) alone (ablation)finetuned Wan2.1 · ConsIDVid-Bench ablation (60-video subset, 50% data)Subject Consistency: 90.83 -> 89.65No clear effect
add view-assisted unposed images (ablation)finetuned Wan2.1 · ConsIDVid-Bench ablation (60-video subset, 50% data)Subject Consistency: 90.83 -> 91.87Improved
add geometry encoder, view-assist and text-visual connector (full model) (ablation)finetuned Wan2.1 · ConsIDVid-Bench ablation (60-video subset, 50% data)Subject Consistency: 91.87 -> 95.13Improved
use T2V instead of I2V conditioning (Wan2.1)video generation · VBench identitySubject Consistency: 91.84 -> 96.72Improved

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