SYNTOLOGY HomeExplorerAtlasCodeMethodologyAboutDevelopersFeedPricing
Paper · 2309.01265 · ICCV · 2023

SOAR: Scene-debiasing Open-set Action Recognition

Junsong Yuan, Yi Wu, Yuanhao Zhai, David Doermann, Ziyi Liu, Zhenyu Wu, Gang Hua, Chunluan Zhou

arXiv · PDF · Open in the Atlas

Code that ran

We lifted 12 functions out of this paper's own repositories and ran 10 of them in a sandbox. "Ran" means the function executed on a synthesized input and returned a value. It is not a reproduction of the paper's results.

RepositoryRoleRan
yhZhai/SOAR canonical 10 of 12
FunctionStatusWhere it lives
anchor Ran yhZhai/SOAR/docs_zh_CN/stat.py
code served (permissive licence) · get_code("d0ddd50f31a39641")
get_indices Ran yhZhai/SOAR/experiments/draw_uncertainty_distribution.py
code served (permissive licence) · get_code("49c3c12dc62fa96b")
get_ood_acc Ran yhZhai/SOAR/experiments/open_set_evaluation.py
code served (permissive licence) · get_code("e06ec986367edd6c")
get_ood_datalist Ran yhZhai/SOAR/experiments/analyze_scene_bias.py
code served (permissive licence) · get_code("ea1109a89f638356")
get_open_maf1 Ran yhZhai/SOAR/experiments/analyze_scene_bias.py
code served (permissive licence) · get_code("8c4fbd5e5329d6c7")
get_stochastic_uncertainty_fn Ran yhZhai/SOAR/experiments/ood_detection.py
code served (permissive licence) · get_code("03feb5deb14d8343")
get_video_name_from_datalist Ran yhZhai/SOAR/experiments/analyze_scene_bias.py
code served (permissive licence) · get_code("227c51480967ab11")
gram_linear Ran yhZhai/SOAR/experiments/compare_feature_similarity.py
code served (permissive licence) · get_code("65c16032c4c2e917")
gram_rbf Ran yhZhai/SOAR/experiments/compare_feature_similarity.py
code served (permissive licence) · get_code("80314ffcc884fcc5")
to_numpy Ran yhZhai/SOAR/experiments/compare_feature_similarity.py
code served (permissive licence) · get_code("3187403d89a3003a")
get_class_names Not yet run yhZhai/SOAR/experiments/draw_uncertainty_distribution.py
code served (permissive licence) · get_code("6d45c7858760b1bb")
get_results Not yet run yhZhai/SOAR/experiments/analyze_recon_uncertainty.py
code served (permissive licence) · get_code("8dd75826e2ff1516")

Repositories linked to this paper

Some links come from the archived Papers with Code dataset (CC BY-SA 4.0): attribution and licence.

Abstract

Deep learning models have a risk of utilizing spurious clues to make predictions, such as recognizing actions based on the background scene. This issue can severely degrade the open-set action recognition performance when the testing samples have different scene distributions from the training samples. To mitigate this problem, we propose a novel method, called Scene-debiasing Open-set Action Recognition (SOAR), which features an adversarial scene reconstruction module and an adaptive adversarial scene classification module. The former prevents the decoder from reconstructing the video background given video features, and thus helps reduce the background information in feature learning. The latter aims to confuse scene type classification given video features, with a specific emphasis on the action foreground, and helps to learn scene-invariant information. In addition, we design an experiment to quantify the scene bias. The results indicate that the current open-set action recognizers are biased toward the scene, and our proposed SOAR method better mitigates such bias. Furthermore, our extensive experiments demonstrate that our method outperforms state-of-the-art methods, and the ablation studies confirm the effectiveness of our proposed modules.

For agents

The same record, over MCP at https://syntology.ai/mcp:

get_harvested_code_for_paper("2309.01265")
get_code_for_paper("2309.01265")
have("2309.01265")

Connect an agent — have() is free.