论文标题
IQIYI提交活动网络挑战2019 Kinetics-700挑战：分层小组关注
iqiyi Submission to ActivityNet Challenge 2019 Kinetics-700 challenge: Hierarchical Group-wise Attention
论文作者
论文摘要
在本报告中，描述了IQIYI提交的方法对活动网络2019 Kinetics-700挑战的任务。模型集合阶段涉及三个模型：TSN，HG-NL和STNET。我们建议用于视频分类的框架级特征聚合的分层群体非本地（HG-NL）模块。标准的非本地（NL）模块在视频分类任务中汇总帧级特征有效，但呈现低参数效率和高计算成本。 HG-NL方法涉及层次结构结构，并生成多个注意图以提高性能。基于该分层组的结构，所提出的方法具有竞争精度，比标准NL的参数更少，计算成本较少。对于ActivityNet 2019 Kinetics-700挑战的任务，在模型集合之后，我们最终在测试集中获得了平均TOP-1和TOP-5错误百分比28.444％。
In this report, the method for the iqiyi submission to the task of ActivityNet 2019 Kinetics-700 challenge is described. Three models are involved in the model ensemble stage: TSN, HG-NL and StNet. We propose the hierarchical group-wise non-local (HG-NL) module for frame-level features aggregation for video classification. The standard non-local (NL) module is effective in aggregating frame-level features on the task of video classification but presents low parameters efficiency and high computational cost. The HG-NL method involves a hierarchical group-wise structure and generates multiple attention maps to enhance performance. Basing on this hierarchical group-wise structure, the proposed method has competitive accuracy, fewer parameters and smaller computational cost than the standard NL. For the task of ActivityNet 2019 Kinetics-700 challenge, after model ensemble, we finally obtain an averaged top-1 and top-5 error percentage 28.444% on the test set.
