论文标题
半监督单视图3D重建通过原型形状先验
Semi-Supervised Single-View 3D Reconstruction via Prototype Shape Priors
论文作者
论文摘要
现有的单视3D重建方法的性能在很大程度上依赖于大规模的3D注释。但是，这种注释繁琐而昂贵。半监督学习是减轻对手动标签需求的另一种方法，但仍未在3D重建中探索。受到半监督图像分类任务的最新成功的启发，我们提出了SSP3D，这是3D重建的半监督框架。特别是，我们引入了一个注意引导的原型形状先验模块，以指导逼真的对象重建。我们进一步介绍了一个歧视者指导的模块，以激励更好的形状产生，并适用于容忍嘈杂的训练样本。在Shapenet基准上，所提出的方法在各种标签率下（即1％，5％，10％和20％）优于先前的监督方法。此外，我们的方法在标记比为10％的标签率下转移到现实世界中的Pix3D数据集时也表现良好。我们还证明了我们的方法可以通过很少的新型监督数据转移到新型类别。流行的Shapenet数据集的实验表明，我们的方法的表现优于零射基线的基线超过12％，我们还进行了严格的消融和分析以验证我们的方法。
The performance of existing single-view 3D reconstruction methods heavily relies on large-scale 3D annotations. However, such annotations are tedious and expensive to collect. Semi-supervised learning serves as an alternative way to mitigate the need for manual labels, but remains unexplored in 3D reconstruction. Inspired by the recent success of semi-supervised image classification tasks, we propose SSP3D, a semi-supervised framework for 3D reconstruction. In particular, we introduce an attention-guided prototype shape prior module for guiding realistic object reconstruction. We further introduce a discriminator-guided module to incentivize better shape generation, as well as a regularizer to tolerate noisy training samples. On the ShapeNet benchmark, the proposed approach outperforms previous supervised methods by clear margins under various labeling ratios, (i.e., 1%, 5% , 10% and 20%). Moreover, our approach also performs well when transferring to real-world Pix3D datasets under labeling ratios of 10%. We also demonstrate our method could transfer to novel categories with few novel supervised data. Experiments on the popular ShapeNet dataset show that our method outperforms the zero-shot baseline by over 12% and we also perform rigorous ablations and analysis to validate our approach.
