Research map / Video / Video generation
LVTINO: LAtent Video consisTency INverse sOlver for High Definition Video Restoration
Alessio Spagnoletti, Andres Almansa, Marcelo Pereyra
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 |
|---|---|---|---|
| use TV3 regularization in the VCM data-consistency step instead of plain prox (ablation) | LVTINO · Adobe240 Problem A (temporal SR x4 + SR x4) | FVMD: 607.5 -> 371.1 | Improved |
| use TV3 regularization in the VCM step instead of plain prox (ablation) | LVTINO · Adobe240 Problem A (temporal SR x4 + SR x4) | PSNR: 22.59 -> 27.25 | Improved |
| use TV3 regularization in the VCM step instead of plain prox (ablation) | LVTINO · Adobe240 Problem A (temporal SR x4 + SR x4) | LPIPS: 0.475 -> 0.249 | Improved |
| add TV3 regularization to the VCM step (Problem A config) (ablation) | LVTINO · Adobe240 Problem B (temporal blur + SR x8) | FVMD: 42.65 -> 51.52 | Hurt |
| add TV3 regularization to the VCM step (Problem A config) (ablation) | LVTINO · Adobe240 Problem B (temporal blur + SR x8) | PSNR: 24.91 -> 23.18 | Hurt |
| add the image consistency prior (ICM) to the VCM-only solver, prox (ablation) | LVTINO-V · Adobe240 Problem A (temporal SR x4 + SR x4) | LPIPS: 0.270 -> 0.475 | Hurt |
| add the image consistency prior (ICM) to the VCM-only solver, prox (ablation) | LVTINO-V · Adobe240 Problem B (temporal blur + SR x8) | LPIPS: 0.375 -> 0.370 | Improved |
| add the image consistency prior (ICM) to the VCM-only solver, prox (ablation) | LVTINO-V · Adobe240 Problem B (temporal blur + SR x8) | FVMD: 31.70 -> 42.65 | Hurt |
| use TV3 instead of prox in VCM-only solver LVTINO-V (ablation) | LVTINO-V · Adobe240 Problem B (temporal blur + SR x8) | FVMD: 31.70 -> 578.0 | Hurt |
| use TV3 instead of prox in VCM-only solver LVTINO-V (ablation) | LVTINO-V · Adobe240 Problem A (temporal SR x4 + SR x4) | FVMD: 425.2 -> 503.3 | Hurt |
| replace VISION-XL with LVTINO (video consistency prior) | LVTINO · Adobe240 Problem A (temporal SR x4 + SR x4) | FVMD: 1141 -> 371.1 | Improved |
| replace VISION-XL with LVTINO | LVTINO · Adobe240 Problem C (temporal SR x8 + SR x8) | FVMD: 1604 -> 602.5 | Improved |
Later papers that compared with this method
1 later accepted paper reports LVTINO as a baseline in their results tables; 1 of them report a better number on most of the results they share with it.
- Accelerating Video Inverse Problem Solvers with Autoregressive Diffusion Models (NeurIPS 2026 · better on 7 of 7 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- VISION-XL — VISION-XL: High Definition Video Inverse Problem Solver using Latent Image Diffusion Models (ICCV 2025) · better on 13 of 18 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
- One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution (NeurIPS 2025 · similarity 0.89)
- Warped Diffusion: Solving Video Inverse Problems with Image Diffusion Models (NeurIPS 2024 · similarity 0.89)
- Solving Inverse Problems with Latent Diffusion Models via Hard Data Consistency (ICLR 2024 · similarity 0.88)
- Solving Video Inverse Problems Using Image Diffusion Models (ICLR 2025 · similarity 0.88)
- Accelerating Video Inverse Problem Solvers with Autoregressive Diffusion Models (NeurIPS 2026 · similarity 0.87)
- Upscale-A-Video: Temporal-Consistent Diffusion Model for Real-World Video Super-Resolution (CVPR 2024 · similarity 0.87)
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.
