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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| ConsID-Gen vs Wan2.2-5B on proprietary subset | I2V generation · ConsIDVid-Bench proprietary subset | I2V Subject consistency: 96.85 -> 98.31 | Improved |
| ConsID-Gen vs Wan2.1-14B on proprietary subset | I2V generation · ConsIDVid-Bench proprietary subset | MEt3R: 0.1572 -> 0.0978 | Improved |
| ConsID-Gen vs Wan2.1-14B on proprietary subset | I2V generation · ConsIDVid-Bench proprietary subset | Chamfer Distance: 0.0866 -> 0.0996 | Hurt |
| ConsID-Gen vs Wan2.1-14B on public subset | I2V generation · ConsIDVid-Bench public subset | MEt3R: 0.0961 -> 0.1321 | Hurt |
| ConsID-Gen vs Wan2.2-5B on public subset | I2V generation · ConsIDVid-Bench public subset | I2V Subject consistency: 98.47 -> 98.14 | Hurt |
| ConsID-Gen vs Wan2.1-14B on public subset | I2V generation · ConsIDVid-Bench public subset | Chamfer Distance: 0.1322 -> 0.1277 | Improved |
| add geometry encoder (VGGT) alone (ablation) | finetuned Wan2.1 · ConsIDVid-Bench ablation (60-video subset, 50% data) | Subject Consistency: 90.83 -> 89.65 | No clear effect |
| add view-assisted unposed images (ablation) | finetuned Wan2.1 · ConsIDVid-Bench ablation (60-video subset, 50% data) | Subject Consistency: 90.83 -> 91.87 | Improved |
| 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 -> 95.13 | Improved |
| use T2V instead of I2V conditioning (Wan2.1) | video generation · VBench identity | Subject Consistency: 91.84 -> 96.72 | Improved |
Closest papers by meaning
- Mask^2DiT: Dual Mask-based Diffusion Transformer for Multi-Scene Long Video Generation (CVPR 2025 · similarity 0.88)
- StreamingT2V: Consistent, Dynamic, and Extendable Long Video Generation from Text (CVPR 2025 · similarity 0.88)
- TIV-Diffusion: Towards Object-Centric Movement for Text-driven Image to Video Generation (AAAI 2025 · similarity 0.88)
- OpenVid-1M: A Large-Scale High-Quality Dataset for Text-to-video Generation (ICLR 2025 · similarity 0.88)
- Encapsulated Composition of Text-to-Image and Text-to-Video Models for High-Quality Video Synthesis (CVPR 2025 · similarity 0.88)
- Snap Video: Scaled Spatiotemporal Transformers for Text-to-Video Synthesis (CVPR 2024 · similarity 0.88)
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.
