Get 3D scene using Segment Anything Model and spherical gradient refinement with SD-MVS
Get 3D scene using Segment Anything Model and spherical gradient refinement with SD-MVS
SD-MVS: Segmentation-Driven Deformation Multi-View Stereo with Spherical Refinement and EM optimization
arXiv paper abstract https://arxiv.org/abs/2401.06385
arXiv PDF paper https://arxiv.org/pdf/2401.06385.pdf
… introduce Segmentation-Driven Deformation Multi-View Stereo (SD-MVS), a method that can effectively tackle challenges in 3D reconstruction of textureless areas.
… 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.
… propose … refinement … that combines spherical coordinates and gradient descent on normals and pixelwise search interval on depths … improving the completeness of reconstructed 3D model.
… adopt the Expectation-Maximization (EM) algorithm to alternately optimize the aggregate matching cost and hyperparameters … mitigating the problem of parameters being excessively dependent on empirical tuning.
… method can achieve state-of-the-art results with less time consumption.
Stay up to date. Subscribe to my posts https://morrislee1234.wixsite.com/website/contact
Web site with my other posts by category https://morrislee1234.wixsite.com/website
LinkedIn https://www.linkedin.com/in/morris-lee-47877b7b
