Recently, 3D scenes parsing with deep learning approaches has been a heating topic. However, current methods with fully-supervised models require manually annotated point-wise supervision which is extremely user-unfriendly and time-consuming to obtain. As such, training 3D scene parsing models with sparse supervision is an intriguing alternative. We term this task as data-efficient 3D scene parsing and propose an effective two-stage framework named VIBUS to resolve it by exploiting the enormous unlabeled points. In the first stage, we perform self-supervised representation learning on unlabeled points with the proposed Viewpoint Bottleneck loss function. The loss function is derived from an information bottleneck objective imposed on scenes under different viewpoints, making the process of representation learning free of degradation and sampling. In the second stage, pseudo labels are harvested from the sparse labels based on uncertainty-spectrum modeling. By combining data-driven uncertainty measures and 3D mesh spectrum measures (derived from normal directions and geodesic distances), a robust local affinity metric is obtained. Finite gamma/beta mixture models are used to decompose category-wise distributions of these measures, leading to automatic selection of thresholds. We evaluate VIBUS on the public benchmark ScanNet and achieve state-of-the-art results on both validation set and online test server. Ablation studies show that both Viewpoint Bottleneck and uncertainty-spectrum modeling bring significant improvements. Codes and models are publicly available at https://github.com/AIR-DISCOVER/VIBUS.
翻译:最近,3D场景与深层学习方法的剖析一直是供暖话题。然而,目前采用完全监督模型的方法需要人工加注的点对点监督,而这种监督非常不方便用户,而且耗费时间才能获得。因此,培训3D场对点分析模型而少监督是一个令人感兴趣的替代办法。我们将此任务称为数据效率3D场对点的剖析,并提议一个名为VIBUS的两阶段有效框架,以利用巨大的未标点加以解决。在第一阶段,我们用拟议的Viewpoint Bottleneck损失功能在未标点上进行自我监督的代表学习。损失功能源自于在不同角度对场景强加的信息瓶颈目标,使代表过程无需退化和取样。在第二阶段,假标签是从基于不确定性光谱模型的稀疏标签中提取的。通过将数据驱动的不确定性模型计量和3Dmelgel(来自正常方向和地理偏差)的频谱计量,在拟议的Vettlenational-commilational Referational Reforational Real-deal developational developmental developmental ress) Abal-Bral press 和Wegradustreval pressmental missueal develildal degradustrismlational messal messal res-bal messal messal messal deal ress press pressal develessal ress ress ress ress 和Webismlationalizlationalmessal deplational ress ressalbress ress lactions ress ress ress ress ressal ress ressal ressal ressal ress ress ress labal labalbalbalbalbalbalbalbal ressal ress res res ress labal labal ress ress ress labal ress ress labal resemememememememememememememememems ress