论文标题
有希望还是难以捉摸？来自现实世界的单个图像的无监督对象分割
Promising or Elusive? Unsupervised Object Segmentation from Real-world Single Images
论文作者
论文摘要
在本文中，我们研究了单个图像中无监督对象分割的问题。我们没有引入新的算法，而是系统地研究现有模型对挑战现实世界图像的有效性。首先，我们引入了四个复杂性因素，以定量测量具有人类注释的数据集中的对象和场景级别的偏差和几何形状的分布。借助这些因素，我们从经验上发现，毫不奇怪的是，现有的无监督模型在现实世界图像中灾难性上未能将通用对象细分，尽管由于它们在众多简单的合成数据集上很容易在综合图像和真实图像之间的物体偏见上很大差异。通过对多个消融现实世界数据集进行广泛的实验，我们最终发现，现实世界图像上现有模型巨大故障的主要因素是外观和几何形状中对象和场景级别偏置的挑战性分布。因此，在现有的无监督模型中引入的归纳偏见几乎无法捕获各种对象分布。我们的研究结果表明，未来的工作应利用网络设计中的更明确的客体偏见。
In this paper, we study the problem of unsupervised object segmentation from single images. We do not introduce a new algorithm, but systematically investigate the effectiveness of existing unsupervised models on challenging real-world images. We firstly introduce four complexity factors to quantitatively measure the distributions of object- and scene-level biases in appearance and geometry for datasets with human annotations. With the aid of these factors, we empirically find that, not surprisingly, existing unsupervised models catastrophically fail to segment generic objects in real-world images, although they can easily achieve excellent performance on numerous simple synthetic datasets, due to the vast gap in objectness biases between synthetic and real images. By conducting extensive experiments on multiple groups of ablated real-world datasets, we ultimately find that the key factors underlying the colossal failure of existing unsupervised models on real-world images are the challenging distributions of object- and scene-level biases in appearance and geometry. Because of this, the inductive biases introduced in existing unsupervised models can hardly capture the diverse object distributions. Our research results suggest that future work should exploit more explicit objectness biases in the network design.
