Awesome-Learning-MVS (Methods and Datasets)

• Learning-based MVS Methods

  • [paper - PAMI](https://ieeexplore.ieee.org/document/9099504)]
  • [paper - *voxel occupancy grids* decoded from 3D Grid or *per-view depth maps* decoded after a projection operation.)
  • [paper - Patch_Similarity_ICCV_2017_supplemental.pdf)] (Note: Learning to measure multi-image patch similiarity, NOT end-to-end learning MVS pipeline)
  • [paper
  • [paper
  • [paper
  • [paper - Based_Multi-View_Stereo_ICCV_2019_supplemental.pdf)] [[Github](https://github.com/callmeray/PointMVSNet)] [[T-PAMI](https://ieeexplore.ieee.org/abstract/document/9076298)] (Point-MVSNet performs multi-view stereo reconstruction in a *coarse-to-fine* fashion, learning to predict the 3D flow of each point to the groundtruth surface based on geometry priors and 2D image appearance cues)
  • [paper
  • [paper
  • [paper
  • [paper - stereo)]
  • [paper
  • [paper - MVSNet)]
  • [paper - MVSNet_Sparse-to-Dense_Multi-View_CVPR_2020_supplemental.pdf)] [[Github](https://github.com/svip-lab/FastMVSNet)]
  • [paper
  • [paper - liujin/REDNet)] [[data](http://gpcv.whu.edu.cn/data/WHU_MVS_Stereo_dataset.html)]
  • [paper - yhw/PVAMVSNet)]
  • [paper - yhw/D2HC-RMVSNet)]
  • [paper - MVSNet)]
  • [paper
  • [paper
  • [paper - RMVSNet_Adaptive_Aggregation_ICCV_2021_supplemental.pdf)] [[Github](https://github.com/QT-Zhu/AA-RMVSNet)]
  • [paper
  • [paper
  • [paper
  • [paper
  • [paper
  • [paper - Based_CVPR_2022_supplemental.pdf)] [[Github](https://github.com/Airobin329/RayMVSNet)]
  • [paper - Parametric_Depth_Distribution_CVPR_2022_supplemental.pdf)]
  • [paper - Aware_CVPR_2022_supplemental.pdf)] [[Github](https://github.com/MegviiRobot/TransMVSNet)]
  • [paper - Net)]
  • [paper - View_Stereo_CVPR_2022_supplemental.pdf)] [[Github](https://github.com/bdwsq1996/Effi-MVS)]
  • [paper - View_CVPR_2022_supplemental.pdf)] [[Github](https://github.com/zhenpeiyang/MVS2D)]
  • [paper
  • [paper - vl/CER-MVS)]
  • [paper
  • [paper
  • [paper - MVSNet)]
  • [paper
  • [paper
  • [paper
  • [paper

• ~~To Be Continued~~

  • [ArXiv
  • [paper
  • [paper
  • [paper
  • [paper - Net)]
  • [paper
  • Robust Methods for Accurate and Efficient 3D Modeling from Unstructured Imagery
  • [paper

• Multi-view Stereo Benchmark

  • [paper
  • [paper
  • paper: [CVPR2014 - 016-0902-9.pdf)] [[website](http://roboimagedata.compute.dtu.dk/?page_id=36)] [[Eval code](https://github.com/Todd-Qi/MVSNet-PyTorch/tree/master/evaluations/dtu)] [[video](https://www.bilibili.com/video/BV1k5411G7NA/)]
  • [paper - ePgl6HF260MGhQX0dCcmdHbFk)] [[website](https://www.tanksandtemples.org/)] [[Github](https://github.com/intel-isl/TanksAndTemples)] [[leaderboard](https://www.tanksandtemples.org/leaderboard/)]
  • [paper - supp.pdf)] [[website](https://www.eth3d.net/)] [[Github](https://github.com/ETH3D)]
  • [paper - Scale_CVPR_2020_supplemental.pdf)] [[Github](https://github.com/YoYo000/BlendedMVS)] [[visual](https://github.com/kwea123/BlendedMVS_scenes)]
  • [paper
  • [paper

• Large-scale Real-world Scenes

  • ImageDataset_SceauxCastle

• Other similar collections

  • Awesome-MVS1
  • Awesome-MVS2
  • PatchMatch Multi-view Stereo
  • multi-view-3d-reconstruction
  • PatchMatch Multi-view Stereo

• Future works (Personal Perspective)

  • GigaMVS