论文标题
学会通过自我统治调整多视图立体声
Learning to Adapt Multi-View Stereo by Self-Supervision
论文作者
论文摘要
来自多个视图的3D场景重建是计算机视觉中的重要经典问题。基于深度学习的方法最近显示出令人印象深刻的重建结果。当培训此类模型时，自我监督的方法是有利的，因为它们不依赖地面真相数据，而这是受监督培训所需的，并且通常很难获得。此外，学到的多视图立体重建易于环境变化，应该坚定地推广到不同的域。我们为多视图立体声提出了一种自适应学习方法，该方法可以训练深层神经网络，以提高对新目标领域的适应性。我们使用模型 - 静态的元学习（MAML）来训练基本参数，而这些参数又可以通过自我监督的训练来适用于新域上的多视图立体声。我们的评估表明，所提出的适应方法有效地学习了新领域中的自我保护的多视图立体声重建。
3D scene reconstruction from multiple views is an important classical problem in computer vision. Deep learning based approaches have recently demonstrated impressive reconstruction results. When training such models, self-supervised methods are favourable since they do not rely on ground truth data which would be needed for supervised training and is often difficult to obtain. Moreover, learned multi-view stereo reconstruction is prone to environment changes and should robustly generalise to different domains. We propose an adaptive learning approach for multi-view stereo which trains a deep neural network for improved adaptability to new target domains. We use model-agnostic meta-learning (MAML) to train base parameters which, in turn, are adapted for multi-view stereo on new domains through self-supervised training. Our evaluations demonstrate that the proposed adaptation method is effective in learning self-supervised multi-view stereo reconstruction in new domains.
