Research map / Video / Video segmentation and tracking
Learning Energy-Based Spatial-Temporal Generative ConvNets for Dynamic Patterns
Jianwen Xie, Song Chun Zhu, Ying Nian Wu
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 |
|---|---|---|---|
| replace Doretto et al. linear dynamic system with the energy-based spatial-temporal ConvNet | dynamic texture synthesis · 17 dynamic textures | SSIM: 0.5602 -> 0.6020 | Improved |
| use the energy-based model instead of HOSVD | dynamic texture synthesis · 17 dynamic textures | SSIM: 0.2577 -> 0.6020 | Improved |
| use the energy-based model instead of FFT-LDS | dynamic texture synthesis · 17 dynamic textures | SSIM: 0.2569 -> 0.6020 | Improved |
| use the energy-based model instead of a 3D MRF with l1 potentials | video recovery · salt and pepper masks | recovery error: 10.1071 -> 5.3048 | Improved |
| use the energy-based model instead of a 3D MRF with l2 potentials | video recovery · salt and pepper masks | recovery error: 12.8272 -> 5.3048 | Improved |
| use the energy-based model instead of MRF-l1 | video recovery · single region mask | recovery error: 12.2856 -> 6.3484 | Improved |
| use the energy-based model instead of MRF-l1 | video recovery · 50% missing frames | recovery error: 15.0085 -> 6.7285 | Improved |
| add spatially fully connected second layer for non-homogeneous textures | Experiment 1 network · textures with structured background | qualitative | Improved |
Closest papers by meaning
- Synthesizing Dynamic Patterns by Spatial-Temporal Generative ConvNet (CVPR 2017 · similarity 0.97)
- Generative Image Dynamics (CVPR 2024 · similarity 0.86)
- GenRec: Unifying Video Generation and Recognition with Diffusion Models (NeurIPS 2024 · similarity 0.85)
- LAMP: Learn A Motion Pattern for Few-Shot Video Generation (CVPR 2024 · similarity 0.85)
- FreeLong: Training-Free Long Video Generation with SpectralBlend Temporal Attention (NeurIPS 2024 · similarity 0.85)
- Autoregressive Video Generation without Vector Quantization (ICLR 2025 · similarity 0.85)
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.
