Research map / Video / Video segmentation and tracking
STCDiT: Spatio-Temporally Consistent Diffusion Transformer for High-Quality Video Super-Resolution
Junyang Chen, Jiangxin Dong, Long Sun, Yixin Yang, Jinshan Pan
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 |
|---|---|---|---|
| add motion-aware VAE reconstruction (ablation) | pretrained video VAE · REDS30 VAE reconstruction | PSNR: 27.22 -> 31.42 | Improved |
| add motion-aware VAE reconstruction (ablation) | pretrained video VAE · REDS30 VAE reconstruction | SSIM: 0.7802 -> 0.8924 | Improved |
| add first-frame anchor interaction in self-attention (ablation) | Base LoRA DiT · Ablation (anchor-frame study) | MUSIQ: 68.30 -> 70.58 | Improved |
| add first-frame anchor interaction in self-attention (ablation) | Base LoRA DiT · Ablation (anchor-frame study) | DOVER: 55.62 -> 58.68 | Improved |
| add ACFM for anchor feature injection (ablation) | Base w/ FF · Ablation (anchor-frame study) | MUSIQ: 70.58 -> 73.24 | Improved |
| add ACFM for anchor feature injection (ablation) | Base w/ FF · Ablation (anchor-frame study) | DOVER: 58.68 -> 59.99 | Improved |
| replace depth-wise conv injection with ACFM (ablation) | Base w/ FF · Ablation (anchor-frame study) | MUSIQ: 71.87 -> 73.24 | Improved |
| replace depth-wise conv injection with ACFM (ablation) | Base w/ FF · Ablation (anchor-frame study) | DOVER: 59.34 -> 59.99 | Improved |
| add anchor-frame feature refinement module (ablation) | Base w/ FF&ACFM · Ablation (anchor-frame study) | MUSIQ: 73.24 -> 73.57 | Improved |
| add anchor-frame feature refinement module (ablation) | Base w/ FF&ACFM · Ablation (anchor-frame study) | DOVER: 59.99 -> 60.81 | Improved |
| let anchor tokens interact with text embeddings in cross-attention (ablation) | STCDiT · Ablation (anchor-frame study) | MUSIQ: 73.57 -> 68.73 | Hurt |
| let anchor tokens interact with text embeddings in cross-attention (ablation) | STCDiT · Ablation (anchor-frame study) | DOVER: 60.81 -> 56.15 | Hurt |
Closest papers by meaning
- Upscale-A-Video: Temporal-Consistent Diffusion Model for Real-World Video Super-Resolution (CVPR 2024 · similarity 0.90)
- Learning Spatial Adaptation and Temporal Coherence in Diffusion Models for Video Super-Resolution (CVPR 2024 · similarity 0.88)
- One-Step Diffusion for Detail-Rich and Temporally Consistent Video Super-Resolution (NeurIPS 2025 · similarity 0.88)
- Effective Diffusion Transformer Architecture for Image Super-Resolution (AAAI 2025 · similarity 0.88)
- DOVE: Efficient One-Step Diffusion Model for Real-World Video Super-Resolution (NeurIPS 2025 · similarity 0.88)
- SkipSR: Faster Super-Resolution with Token Skipping (NeurIPS 2026 · 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.
