Research map / Video / Video segmentation and tracking
SME-Net: Sparse Motion Estimation for Parametric Video Prediction Through Reinforcement Learning
Yung-Han Ho, Chuan-Yuan Cho, Wen-Hsiao Peng, Guo-Lun Jin
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 |
|---|---|---|---|
| increase number of critical pixels at test time from 20 to 100 (ablation) | SME-Net · CIF one-step | PSNR and SSIM | Improved |
| increase number of critical pixels at test time (Coastguard sequence) (ablation) | SME-Net · CIF Coastguard one-step | SSIM | Hurt |
| use sparse motion field with POBMC instead of dense flow (DVF) | SME-Net vs DVF · CIF one-step | PSNR: 22.75 -> 27.90 | Improved |
| use sparse motion field instead of dense flow (DVF) | SME-Net vs DVF · CaltechPed one-step | MSE (1e-3): 6.65 -> 2.65 | Improved |
| use sparse motion field instead of dense flow (DVF) | SME-Net vs DVF · UCF101 one-step | PSNR: 31.54 -> 30.80 | Hurt |
| use sparse motion model instead of MCNet generative model | SME-Net vs MCNet · CaltechPed multi-step | FVD: 148 -> 132 | Improved |
| use sparse motion model instead of BeyondMSE | SME-Net vs BeyondMSE · CaltechPed multi-step | FVD: 846 -> 132 | Improved |
| use SME-Net instead of MCNet | SME-Net vs MCNet · CIF one-step | PSNR: 28.34 -> 27.90 | Hurt |
| use SME-Net instead of BeyondMSE with much smaller model | SME-Net vs BeyondMSE · UCF101 one-step | PSNR: 26.7 -> 30.80 | Improved |
Closest papers by meaning
- The Utility of Sparse Representations for Control in Reinforcement Learning (AAAI 2019 · similarity 0.84)
- The State of Sparse Training in Deep Reinforcement Learning (ICML 2022 · similarity 0.83)
- FFNet: Video Fast-Forwarding via Reinforcement Learning (CVPR 2018 · similarity 0.83)
- Towards Principled Representation Learning from Videos for Reinforcement Learning (ICLR 2024 · similarity 0.83)
- Learning Video-Conditioned Policy on Unlabelled Data with Joint Embedding Predictive Transformer (ICLR 2025 · similarity 0.83)
- SDC-Net: Video prediction using spatially-displaced convolution (ECCV 2018 · similarity 0.83)
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.
