Research map / Video / Action recognition
Temporal Interlacing Network
Hao Shao, Shengju Qian, Yu Liu
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 fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · Something-Something V1 | Acc1: 45.5 -> 47.0 | Improved |
| Replace fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · Something-Something V2 | Acc1: 59.2 -> 60.0 | Improved |
| Replace fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · Jester | Acc1: 96.2 -> 96.7 | Improved |
| Replace fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · UCF101 | Acc1: 92.6 -> 93.6 | Improved |
| Replace fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · HMDB51 | Acc1: 69.5 -> 72.0 | Improved |
| Replace fixed temporal shifting with learned temporal interlacing | ResNet-50 video recognition · Kinetics-600 | Acc1: 69.5 -> 70.4 | Improved |
| Use 2 groups rather than 1 or 4 for shifted channels (ablation) | TIN · Something-Something V1 | Val Top-1: 42.0 -> 42.9; 42.4 -> 42.9 | Improved |
| Add reverse offsets to the 2-group configuration (ablation) | TIN · Something-Something V1 | Val Top-1: 41.1 -> 42.9 | Improved |
| Add reverse offsets to the 4-group configuration (ablation) | TIN · Something-Something V1 | Val Top-1: 42.6 -> 42.4 | Hurt |
Closest papers by meaning
- TAda! Temporally-Adaptive Convolutions for Video Understanding (ICLR 2022 · similarity 0.85)
- Across Scales & Across Dimensions: Temporal Super-Resolution using Deep Internal Learning (ECCV 2020 · similarity 0.85)
- Spatio-Temporal Channel Correlation Networks for Action Classification (ECCV 2018 · similarity 0.84)
- StNet: Local and Global Spatial-Temporal Modeling for Action Recognition (AAAI 2019 · similarity 0.84)
- TDN: Temporal Difference Networks for Efficient Action Recognition (CVPR 2021 · similarity 0.84)
- Enhancing Temporal Understanding in Video-LLMs through Stacked Temporal Attention in Vision Encoders (NeurIPS 2025 · 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.
