论文标题
转向老师进行时间戳监督的时间动作细分
Turning to a Teacher for Timestamp Supervised Temporal Action Segmentation
论文作者
论文摘要
视频中的时间动作细分最近引起了很多关注。时间戳监督是完成此任务的一种经济高效的方式。为了获得更多信息以优化模型，现有方法生成的伪框架根据分割模型的输出和时间戳注释对迭代标签进行了迭代标签。但是，这种做法可能会在训练过程中引入噪声和振荡，并导致性能变性。为了解决这个问题，我们通过引入与分割模型平行的教师模型来帮助稳定模型优化的过程，为时间戳监督的时间术及时间动作细分提出了新的框架。教师模型可以看作是分割模型的合奏，这有助于抑制噪声并提高伪标签的稳定性。我们进一步引入了一个分段平滑的损失，该损失更加集中和凝聚力，以实现行动实例中预测概率的平稳过渡。三个数据集的实验表明，我们的方法的表现优于最新方法，并且以较低的注释成本与完全监督的方法相当地执行。
Temporal action segmentation in videos has drawn much attention recently. Timestamp supervision is a cost-effective way for this task. To obtain more information to optimize the model, the existing method generated pseudo frame-wise labels iteratively based on the output of a segmentation model and the timestamp annotations. However, this practice may introduce noise and oscillation during the training, and lead to performance degeneration. To address this problem, we propose a new framework for timestamp supervised temporal action segmentation by introducing a teacher model parallel to the segmentation model to help stabilize the process of model optimization. The teacher model can be seen as an ensemble of the segmentation model, which helps to suppress the noise and to improve the stability of pseudo labels. We further introduce a segmentally smoothing loss, which is more focused and cohesive, to enforce the smooth transition of the predicted probabilities within action instances. The experiments on three datasets show that our method outperforms the state-of-the-art method and performs comparably against the fully-supervised methods at a much lower annotation cost.
