DeepMVS: Learning Multi-View Stereopsis


Po-Han Huang1Kevin Matzen2Johannes Kopf2Narendra Ahuja1Jia-Bin Huang3

1University of Illinois, Urbana Champaign   2Facebook   3Virginia Tech


Abstract

We present DeepMVS, a deep convolutional neural network (ConvNet) for multi-view stereo reconstruction. Taking an arbitrary number of posed images as input, we first produce a set of plane-sweep volumes and use the proposed DeepMVS network to predict high-quality disparity maps. The key contributions that enable these results are (1) supervised training on a photorealistic synthetic dataset, (2) an effective method for aggregating information across a set of unordered images, and (3) integrating multi-layer feature activations from the pre-trained VGG-19 network. We validate the efficacy of DeepMVS using the ETH3D Benchmark. Our results show that DeepMVS compares favorably against state-of-the-art conventional MVS algorithms and other CNN based methods, particularly for near-textureless regions and thin structures.

BibTeX

If you use our code or data, please cite:

    @inproceedings{DeepMVS,
      author       = "Huang, Po-Han and Matzen, Kevin and Kopf, Johannes and Ahuja, Narendra and Huang, Jia-Bin",
      title        = "DeepMVS: Learning Multi-View Stereopsis",
      booktitle    = "IEEE Conference on Computer Vision and Pattern Recognition (CVPR)",
      year         = "2018"
    }