SD-MVS: Segmentation-Driven Deformation Multi-View Stereo with Spherical Refinement and EM optimization
It addresses challenges in multi-view stereo reconstruction for textureless scenes, offering an incremental improvement over existing methods.
The paper tackles 3D reconstruction of textureless areas by introducing SD-MVS, which uses segmentation for deformation and spherical refinement, achieving state-of-the-art results with less time consumption on benchmarks like ETH3D and Tanks and Temples.
In this paper, we introduce Segmentation-Driven Deformation Multi-View Stereo (SD-MVS), a method that can effectively tackle challenges in 3D reconstruction of textureless areas. We are the first to adopt the Segment Anything Model (SAM) to distinguish semantic instances in scenes and further leverage these constraints for pixelwise patch deformation on both matching cost and propagation. Concurrently, we propose a unique refinement strategy that combines spherical coordinates and gradient descent on normals and pixelwise search interval on depths, significantly improving the completeness of reconstructed 3D model. Furthermore, we adopt the Expectation-Maximization (EM) algorithm to alternately optimize the aggregate matching cost and hyperparameters, effectively mitigating the problem of parameters being excessively dependent on empirical tuning. Evaluations on the ETH3D high-resolution multi-view stereo benchmark and the Tanks and Temples dataset demonstrate that our method can achieve state-of-the-art results with less time consumption.