Research map / Video / Action recognition
SeCo: Exploring Sequence Supervision for Unsupervised Representation Learning
Ting Yao, Yiheng Zhang, Zhaofan Qiu, Yingwei Pan, Tao Mei
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 intra-frame discrimination to inter-frame discrimination (ablation) | SeCo · Kinetics400 validation | Top-1: 58.97 -> 60.74 | Improved |
| Add intra-frame discrimination to inter-frame discrimination (ablation) | SeCo · ActivityNet validation | Top-1: 66.69 -> 68.31 | Improved |
| Add intra-frame discrimination to inter-frame discrimination (ablation) | SeCo · OTB-100 | Precision: 67.92 -> 70.29 | Improved |
| Add temporal order validation to inter-frame and intra-frame discrimination (ablation) | SeCo · Kinetics400 validation | Top-1: 60.74 -> 61.91 | Improved |
| Add temporal order validation to inter-frame and intra-frame discrimination (ablation) | SeCo · ActivityNet validation | Top-1: 68.31 -> 68.55 | Improved |
| Add temporal order validation to inter-frame and intra-frame discrimination (ablation) | SeCo · OTB-100 | Success: 50.48 -> 51.78 | Improved |
| Pre-train with SeCo instead of supervised ImageNet pre-training | ResNet50 fine-tuned with TSN · UCF101 | Top-1: 85.30 -> 88.26 | Improved |
| Pre-train with SeCo instead of supervised ImageNet pre-training | ResNet50 fine-tuned with TSN · HMDB51 | Top-1: 49.08 -> 55.55 | Improved |
| Use SeCo pre-training instead of VIE-Full | ResNet50 fine-tuned with TSN · UCF101 | Top-1: 80.40 -> 88.26 | Improved |
| Use SeCo pre-training instead of VIE-Full | ResNet50 fine-tuned with TSN · HMDB51 | Top-1: 52.50 -> 55.55 | Improved |
| Use SeCo (ResNet18) instead of VIE-Full | Kinetics400 representation learning · Kinetics400 validation | Top-1: 48.53 -> 50.81 | Improved |
Later papers that compared with this method
6 later accepted papers report SeCo as a baseline in their results tables; 5 of them report a better number on most of the results they share with it.
- Self-Supervised Representation Learning Framework for Remote Physiological Measurement Using Spatiotemporal Augmentation Loss (AAAI 2022 · better on 3 of 3 shared results)
- Motion-Focused Contrastive Learning of Video Representations (ICCV 2021 · better on 5 of 5 shared results)
- Spatiotemporal Contrastive Video Representation Learning (CVPR 2021 · better on 1 of 1 shared result)
- Contrast and Order Representations for Video Self-Supervised Learning (ICCV 2021 · better on 1 of 1 shared result)
- Rethinking Self-Supervised Correspondence Learning: A Video Frame-Level Similarity Perspective (ICCV 2021 · better on 1 of 1 shared result)
- Contrastive Learning of Image Representations With Cross-Video Cycle-Consistency (ICCV 2021 · better on 0 of 1 shared result)
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Semi-Supervised Action Recognition With Temporal Contrastive Learning (CVPR 2021 · similarity 0.89)
- RSPNet: Relative Speed Perception for Unsupervised Video Representation Learning (AAAI 2021 · similarity 0.87)
- Shuffle and Learn: Unsupervised Learning Using Temporal Order Verification (ECCV 2016 · similarity 0.87)
- Frame-Wise Action Representations for Long Videos via Sequence Contrastive Learning (CVPR 2022 · similarity 0.87)
- Unsupervised Representation Learning by Sorting Sequences (ICCV 2017 · similarity 0.87)
- Fine-Grained Temporal Contrastive Learning for Weakly-Supervised Temporal Action Localization (CVPR 2022 · 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.
