Recent state-of-the-art method FlexMatch firstly demonstrated that correctly estimating learning status is crucial for semi-supervised learning (SSL). However, the estimation method proposed by FlexMatch does not take into account imbalanced data, which is the common case for 3D semi-supervised learning. To address this problem, we practically demonstrate that unlabeled data class-level confidence can represent the learning status in the 3D imbalanced dataset. Based on this finding, we present a novel class-level confidence based 3D SSL method. Firstly, a dynamic thresholding strategy is proposed to utilize more unlabeled data, especially for low learning status classes. Then, a re-sampling strategy is designed to avoid biasing toward high learning status classes, which dynamically changes the sampling probability of each class. To show the effectiveness of our method in 3D SSL tasks, we conduct extensive experiments on 3D SSL classification and detection tasks. Our method significantly outperforms state-of-the-art counterparts for both 3D SSL classification and detection tasks in all datasets.
翻译:最近最先进的方法 FlexMatch 首先显示,正确估计学习状况对于半监督学习(SSL)至关重要。 但是, FlexMatch 提议的估算方法没有考虑到不平衡的数据,这是3D半监督学习的常见情况。 为了解决这个问题,我们实际表明,未标记的数据等级信任度可以代表3D不平衡数据集的学习状态。基于这一发现,我们提出了一个基于 3D SSL 的新型等级信任度方法。首先,提出了一种动态阈值战略,以更多地使用未标记的数据,特别是低学习状态类。然后,重新抽样战略旨在避免偏向高学习状态类,这动态地改变了每个类的抽样概率。为了显示我们3D SSL 任务的方法的有效性,我们进行了关于 3D SSL 分类和检测任务的广泛实验。我们的方法大大优于所有数据集中3D SSL 分类和检测任务的最新对应方。