/Awesome-Learning-MVS-depth

A list of awesome learning-based multi-view stereo papers

Awesome-Learning-MVS (Methods and Datasets)

Other similar collections

Learning-based MVS Methods

  1. Volumetric methods (SurfaceNet)
  2. Depthmap based methods (MVSNet/R-MVSNet and so on)

ICCV2017

  • SurfaceNet: An End-to-end 3D Neural Network for Multiview Stereopsis [paper] [Github] [T-PAMI]
  • Learning a Multi-View Stereo Machine [paper] (LSMs can produce two kinds of outputs - voxel occupancy grids decoded from 3D Grid or per-view depth maps decoded after a projection operation.)
  • Learned Multi-Patch Similarity [paper] [supp] (Note: Learning to measure multi-image patch similiarity, NOT end-to-end learning MVS pipeline)

CVPR2018

ECCV2018

CVPR2019

  • Recurrent MVSNet for High-resolution Multi-view Stereo Depth Inference [paper] [supp] [Github]

ICCV2019

  • Point-Based Multi-View Stereo Network [paper] [supp] [Github] [T-PAMI] (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)
  • P-MVSNet: Learning Patch-wise Matching Confidence Aggregation for Multi-view Stereo [paper]
  • MVSCRF: Learning Multi-view Stereo with Conditional Random Fields [paper]

AAAI2020

  • Learning Inverse Depth Regression for Multi-View Stereo with Correlation Cost Volume [paper] [Github]

CVPR2020

  • Cascade Cost Volume for High-Resolutoin Multi-View Stereo and Stereo Matching [paper] [Github] [WeChat article]

  • Deep Stereo using Adaptive Thin Volume Representation with Uncertainty Awareness [paper] [supp] [Github]

  • Cost Volume Pyramid Based Depth Inference for Multi-View Stereo [paper] [supp] [Github]

  • Fast-MVSNet: Sparse-to-Dense Multi-View Stereo with Learned Propagation and Gauss-Newton Refinement [paper] [supp] [Github] [WeChat article]

  • Attention-Aware Multi-View Stereo [paper]

  • A Novel Recurrent Encoder-Decoder Structure for Large-Scale Multi-view Stereo Reconstruction from An Open Aerial Dataset [paper] [Github] [data]

ECCV2020

  • Pyramid Multi-view Stereo Net with Self-adaptive View aggregation [paper] [Github]
  • Dense Hybird Recurrent Multi-view Stereo Net with Dynamic Consistency Checking [paper] [Github]

BMVC2020

  • Visibility-aware Multi-view Stereo Network [paper] [Github]

WACV2021

  • Long-range Attention Network for Multi-View Stereo [paper]

CVPR2021

  • PatchmatchNet: Learned Multi-View Patchmatch Stereo [paper] [Github]

ICCV2021

  • AA-RMVSNet: Adaptive Aggregation Recurrent Multi-view Stereo Network [paper] [supp] [Github]
  • EPP-MVSNet: Epipolar-Assembling Based Depth Prediction for Multi-View Stereo [paper] [Github]
  • Just a Few Points are All You Need for Multi-view Stereo: A Novel Semi-supervised Learning Method for Multi-view Stereo [paper] [supp]

3DV 2021

CVPR 2022

  • Generalized Binary Search Network for Highly-Efficient Multi-View Stereo [paper] [Github]
  • Rethinking Depth Estimation for Multi-View Stereo: A Unified Representation and Focal Loss [paper] [Github]
  • IterMVS: Iterative Probability Estimation for Efficient Multi-View Stereo [paper] [Github]
  • MVS2D: Efficient Multi-view Stereo via Attention-Driven 2D Convolutions [paper] [Github]

Journal Paper

  • MVSNet++: Learning Depth-Based Attention Pyramid Features for Multi-View Stereo. IEEE TIP [paper]
  • HighRes-MVSNet: A Fast Multi-View Stereo Network for Dense 3D Reconstruction From High-Resolution Images. IEEE Access [paper]
  • AACVP-MVSNet: Attention-aware cost volume pyramid based multi-view stereo network for 3D reconstruction. ISPRS Journal of Photogrammetry and Remote Sensing [paper] [Github]

Survey Paper

  • A Survey on Deep Learning Techniques for Stereo-based Depth Estimation. IEEE T-PAMI [ArXiv] [IEEE Xplore]
  • Deep Learning for Multi-view Stereo via Plane Sweep: A Survey [paper]
  • Multi-view stereo in the Deep Learning Era: A comprehensive revfiew [paper]

ArXiv Paper

  • PVSNet: Pixelwise Visibility-Aware Multi-View Stereo Network [paper]
  • DDR-Net: Learning Multi-Stage Multi-View Stereo With Dynamic Depth Range [paper] [Github]
  • Non-local Recurrent Regularization Networks for Multi-view Stereo [paper]
  • TransMVSNet: Global Context-aware Multi-view Stereo Network with Transformers [paper] [Github]

Multi-view Stereo Benchmark

  • Middlebury [CVPR06']

    • A Comparison and Evaluation of Multi-View Stereo Reconstruction Algorithms [paper] [website]

  • EPFL [CVPR08']

    • On Benchmarking Camera Calibration and Multi-View Stereo for High Resolution Imagery [paper]

  • DTU [CVPR2014, IJCV2016]

  • Tanks and Temples [ACM ToG2017]

  • ETH3D [CVPR2017]

    • A Multi-View Stereo Benchmark with High-Resolution Images and Multi-Camera Videos [paper] [supp] [website] [Github]

  • BlendedMVS [CVPR2020]

  • GigaMVS [T-PAMI2021]

    • GigaMVS: A Benchmark for Ultra-large-scale Gigapixel-level 3D Reconstruction [paper] [website]

  • Multi-sensor large-scale dataset for multi-view 3D reconstruction [ArXiv2022]

    • Multi-sensor large-scale dataset for multi-view 3D reconstruction [paper] [website]

Large-scale Real-world Scenes

  1. Chinese Style Architectures
  1. Western Style Architectures
  1. Aerial Dataset

Update: WeChat Group