This paper introduces a learnable Deformable Hypothesis Sampler (DeformSampler) to address the challenging issue of noisy depth estimation for accurate PatchMatch Multi-View Stereo (MVS). We observe that the heuristic depth hypothesis sampling modes employed by PatchMatch MVS solvers are insensitive to (i) the piece-wise smooth distribution of depths across the object surface, and (ii) the implicit multi-modal distribution of depth prediction probabilities along the ray direction on the surface points. Accordingly, we develop DeformSampler to learn distribution-sensitive sample spaces to (i) propagate depths consistent with the scene's geometry across the object surface, and (ii) fit a Laplace Mixture model that approaches the point-wise probabilities distribution of the actual depths along the ray direction. We integrate DeformSampler into a learnable PatchMatch MVS system to enhance depth estimation in challenging areas, such as piece-wise discontinuous surface boundaries and weakly-textured regions. Experimental results on DTU and Tanks \& Temples datasets demonstrate its superior performance and generalization capabilities compared to state-of-the-art competitors. Code is available at https://github.com/Geo-Tell/DS-PMNet.

翻译：本文提出一种可学习的形变假设采样器（Deformable Hypothesis Sampler，简称DeformSampler），以解决精确PatchMatch多视图立体匹配（MVS）中深度估计噪声大的挑战性问题。我们观察到，PatchMatch MVS求解器采用的启发式深度假设采样模式对于以下两点不敏感：（i）物体表面深度的分段平滑分布，以及（ii）表面点沿视线方向的深度预测概率的隐式多模态分布。为此，我们开发DeformSampler来学习对分布敏感的采样空间，从而（i）传播与场景几何一致的物体表面深度，以及（ii）拟合拉普拉斯混合模型以逼近沿视线方向实际深度的逐点概率分布。我们将DeformSampler集成到可学习的PatchMatch MVS系统中，以增强在分段不连续表面边界和弱纹理区域等挑战性区域的深度估计。在DTU和Tanks & Temples数据集上的实验结果表明，与最先进的竞争对手相比，该方法具有优越的性能和泛化能力。代码见https://github.com/Geo-Tell/DS-PMNet。