Research map / Video / Video segmentation and tracking
Video Autoencoder: Self-Supervised Disentanglement of Static 3D Structure and Motion
Zihang Lai, Sifei Liu, Alexei A. Efros, Xiaolong Wang
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 64x64 voxel spatial resolution instead of 32x32 (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 20.23 -> 23.21 | Improved |
| use 64x64 voxel spatial resolution instead of 128x128 (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 22.39 -> 23.21 | Improved |
| use voxel depth 32 instead of 16 (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 21.92 -> 23.21 | Improved |
| use voxel depth 32 instead of 64 (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 22.05 -> 23.21 | Improved |
| train on 6-frame clips instead of 3-frame clips (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 22.52 -> 23.21 | Improved |
| train on 6-frame clips instead of 10-frame clips (ablation) | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 21.54 -> 23.21 | Improved |
| learn pose jointly with voxels without camera supervision, versus SfMLearner-style depth warping | Video Autoencoder · RealEstate10K ATE | ATE mean: 0.048 -> 0.017 | Improved |
| self-supervised pose and voxels versus SynSin trained with ground-truth cameras | Video Autoencoder · RealEstate10K novel view synthesis | PSNR: 22.31 -> 23.21 | Mixed |
Closest papers by meaning
- Learning Disentangled Representations of Videos with Missing Data (NeurIPS 2020 · similarity 0.89)
- Learning to Decompose and Disentangle Representations for Video Prediction (NeurIPS 2018 · similarity 0.89)
- Unsupervised Learning of Disentangled Representations from Video (NeurIPS 2017 · similarity 0.87)
- Self-Supervised 3D Human Pose Estimation via Part Guided Novel Image Synthesis (CVPR 2020 · similarity 0.87)
- Autodecoding Latent 3D Diffusion Models (NeurIPS 2023 · similarity 0.87)
- S3VAE: Self-Supervised Sequential VAE for Representation Disentanglement and Data Generation (CVPR 2020 · 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.
