An Empirical Study of End-to-End Temporal Action Detection
Open accessThe DOI matches its Crossref record; 15 of the 15 references it cites were found in a registry or an index.
Abstract
Temporal action detection (TAD) is an important yet challenging task in video understanding. It aims to simultaneously predict the semantic label and the temporal interval of every action instance in an untrimmed video. Rather than end-to-end learning, most existing methods adopt a head-only learning paradigm, where the video encoder is pre-trained for action classification, and only the detection head upon the encoder is optimized for TAD. The effect of end-to-end learning is not systematically evaluated. Besides, there lacks an in-depth study on the efficiency-accuracy trade-off in end-to-end TAD. In this paper, we present an empirical study of end-to-end temporal action detection. We validate the advantage of end-to-end learning over head-only learning and observe up to 11% performance improvement. Besides, we study the effects of multiple design choices that affect the TAD performance and speed, including detection head, video encoder, and resolution of input videos. Based on the findings, we build a mid-resolution baseline detector, which achieves the state-of-the-art performance of end-to-end methods while running more than 4×faster. We hope that this paper can serve as a guide for end-to-end learning and inspire future research in this field. Code and models are available at https://github.com/xlliu7/E2E-TAD.
Overview
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
The claims and statistics of this paper haven’t been checked yet.
Free: it reads the open-access text and checks the first 5 claims that cite a source we can check. Create a free account to check every claim.
Self-check
Percentages, sums and reported differences re-computed from the paper’s own numbers.
The self-check needs the paper’s full text.
Citations
Cites
Adam: A Method for Stochastic Optimization
Kingma, D. P., & Ba, J. (2014) · arXiv · doi:10.48550/arxiv.1412.6980 · Cited by 98,946
Faster R-CNN: Towards Real-Time Object Detection with Region Proposal Networks
Ren, S., He, K., Girshick, R., & Sun, J. (2017) · IEEE Transactions on Pattern Analysis and Machine Intelligence · doi:10.1109/tpami.2016.2577031 · Cited by 73,472
End-to-End Object Detection with Transformers
Carion, N., Massa, F., Synnaeve, G., Usunier, N., Kirillov, A., & Zagoruyko, S. (2020) · Lecture Notes in Computer Science · doi:10.1007/978-3-030-58452-8_13 · Cited by 18,285
Quo Vadis, Action Recognition? A New Model and the Kinetics Dataset
Carreira, J., & Zisserman, A. (2017) · 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2017.502 · Cited by 9,570
FCOS: Fully Convolutional One-Stage Object Detection
Tian, Z., Shen, C., Chen, H., & He, T. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00972 · Cited by 7,013
SlowFast Networks for Video Recognition
Feichtenhofer, C., Fan, H., Malik, J., & He, K. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00630 · Cited by 4,007
A Closer Look at Spatiotemporal Convolutions for Action Recognition
Tran, D., Wang, H., Torresani, L., Ray, J., LeCun, Y., & Paluri, M. (2017) · arXiv · doi:10.48550/arxiv.1711.11248 · Cited by 3,664 · cited 2×
“It follows recent works [32, 41] to apply 1D and 2D convolution iteratively.”
MethodActivityNet: A large-scale video benchmark for human activity understanding
Heilbron, F. C., Escorcia, V., Ghanem, B., & Niebles, J. C. (2015) · 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2015.7298698 · Cited by 2,578
Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
Goyal, P., Dollár, P., Girshick, R., Noordhuis, P., Wesolowski, L., Kyrola, A., Tulloch, A., Jia, Y., & He, K. (2017) · arXiv · doi:10.48550/arxiv.1706.02677 · Cited by 2,294 · cited 1×
“We change the learning rate following the linear scaling rule [5] when changing the batch size.”
MethodTemporal Action Detection with Structured Segment Networks
Zhao, Y., Xiong, Y., Wang, L., Wu, Z., Tang, X., & Lin, D. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.317 · Cited by 892
BMN: Boundary-Matching Network for Temporal Action Proposal Generation
Lin, T., Liu, X., Li, X., Ding, E., & Wen, S. (2019) · 2019 IEEE/CVF International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2019.00399 · Cited by 694 · cited 3×
“On ActivityNet, we follow previous works [18, 20, 22, 28, 47, 48, 54] to perform classagnostic localization and use the video-level classification labels from [58].”
BackgroundMethodSelf-citationEnd-to-End Learning of Action Detection from Frame Glimpses in Videos
Yeung, S., Russakovsky, O., Mori, G., & Fei-Fei, L. (2016) · 2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2016.293 · Cited by 585
TURN TAP: Temporal Unit Regression Network for Temporal Action Proposals
Gao, J., Yang, Z., Sun, C., Chen, K., & Nevatia, R. (2017) · 2017 IEEE International Conference on Computer Vision (ICCV) · doi:10.1109/iccv.2017.392 · Cited by 498 · cited 1×
“In these methods, anchors are generated by uniform sampling [3, 7, 8, 11, 35, 45], grouping potential action boundaries [22, 34, 56, 57], or a combination of the them [10, 27].”
MethodGaussian Temporal Awareness Networks for Action Localization
Long, F., Yao, T., Qiu, Z., Tian, X., Luo, J., & Mei, T. (2019) · 2019 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR) · doi:10.1109/cvpr.2019.00043 · Cited by 380
End-to-End, Single-Stream Temporal Action Detection in Untrimmed Videos
Buch, S., Escorcia, V., Ghanem, B., & Niebles, J. C. (2017) · Procedings of the British Machine Vision Conference 2017 · doi:10.5244/c.31.93 · Cited by 240 · cited 2×
“Anchor-free methods [2,18,21,34,52] directly predict the action class for each frame in the video.”
Method
Cited by
End-to-End Temporal Action Detection With Transformer
Liu, X., Wang, Q., Hu, Y., Tang, X., Zhang, S., Bai, S., & Bai, X. (2022) · IEEE Transactions on Image Processing · doi:10.1109/tip.2022.3195321 · Cited by 282
Self-citation(2023) · doi:10.1109/cvpr52729.2023.01808 · Cited by 226
Cai, Z., Ghosh, S., Dhall, A., Gedeon, T., Stefanov, K., & Hayat, M. (2023) · Computer Vision and Image Understanding · doi:10.1016/j.cviu.2023.103818 · Cited by 63
RefineTAD: Learning Proposal-free Refinement for Temporal Action Detection
Feng, Y., Zhang, Z., Quan, R., Wang, L., & Qin, J. (2023) · Proceedings of the 31st ACM International Conference on Multimedia · doi:10.1145/3581783.3611872 · Cited by 8
Metadata
- Authors
- Liu, X., Bai, S., & Bai, X.
- Venue
- 2022 IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR)
- Publisher
- IEEE
- Published
- 2022-06
- Pages
- 19978-19987
- ISSN
- 1063-6919 (ISSN-L)
- OpenAlex
- W4313050661
- Cited by
- 73
- References
- 15
- FWCI
- 6.18
- Type
- conference paper
- Language
- en
- Open access
- green
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.