Research map / Video / Video segmentation and tracking
Two-Stream Action Recognition-Oriented Video Super-Resolution
Haochen Zhang, Dong Liu, Zhiwei Xiong
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 weighted MSE (flow-based weights) instead of MSE (ablation) | VDSR SoSR · HMDB51 TSN spatial SoSR ablation | accuracy (%): 46.6 -> 47.91 | Improved |
| add feature loss (ablation) | VDSR SoSR · HMDB51 TSN spatial SoSR ablation | accuracy (%): 47.91 -> 50.39 | Improved |
| switch network from VDSR to ESRGAN structure (ablation) | SoSR · HMDB51 TSN spatial SoSR ablation | accuracy (%): 50.39 -> 52.55 | Improved |
| add adversarial loss (ablation) | ESRGAN SoSR with WMSE · HMDB51 TSN spatial SoSR ablation | accuracy (%): 52.55 -> 53.59 | Improved |
| use WMSE instead of MSE (ablation) | ESRGAN SoSR with feature and adversarial losses · HMDB51 TSN spatial SoSR ablation | accuracy (%): 52.48 -> 53.59 | Improved |
| add warp loss (ablation) | VDSR ToSR · HMDB51 TSN temporal ToSR ablation | accuracy (%): 55.1 -> 58.76 | Improved |
| add warp loss (ablation) | VSR-DUF-16 ToSR · HMDB51 TSN temporal ToSR ablation | accuracy (%): 59.48 -> 61.5 | Improved |
| use multi-frame VSR-DUF-16 instead of single-frame VDSR (ablation) | ToSR · HMDB51 TSN temporal ToSR ablation | accuracy (%): 55.1 -> 59.48 | Improved |
| use perceptual-loss SR (ESRGAN) instead of MSE-only SR (VDSR) for temporal stream | single-image SR · HMDB51 TSN temporal stream (4x SR) | accuracy (%): 55.1 -> 51.5 | Hurt |
| use perceptual-loss SR (ESRGAN) instead of VDSR for spatial stream | single-image SR · HMDB51 TSN spatial stream (4x SR) | accuracy (%): 46.6 -> 52.48 | Improved |
| combine SoSR and ToSR streams | action recognition on SR video · HMDB51 TSN fusion (4x SR) | accuracy (%): 63.53 -> 68.3 | Improved |
| combine SoSR and ToSR streams | action recognition on SR video · UCF101 TSN fusion (4x SR) | accuracy (%): 87.87 -> 92.13 | Improved |
Closest papers by meaning
- Learning SpatioTemporal and Motion Features in a Unified 2D Network for Action Recognition (TPAMI 2022 · similarity 0.88)
- StNet: Local and Global Spatial-Temporal Modeling for Action Recognition (AAAI 2019 · similarity 0.88)
- Spatiotemporal Residual Networks for Video Action Recognition (NeurIPS 2016 · similarity 0.88)
- Convolutional Two-Stream Network Fusion for Video Action Recognition (CVPR 2016 · similarity 0.88)
- Two-Streams: Dark and Light Networks with Graph Convolution for Action Recognition from Dark Videos (Student Abstract) (AAAI 2023 · similarity 0.87)
- Spatiotemporal Pyramid Network for Video Action Recognition (CVPR 2017 · 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.
