Designing deep networks for surface normal estimation

@article{Wang2015DesigningDN,
  title={Designing deep networks for surface normal estimation},
  author={X. Wang and David F. Fouhey and A. Gupta},
  journal={2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)},
  year={2015},
  pages={539-547}
}
  • X. Wang, David F. Fouhey, A. Gupta
  • Published 2015
  • Computer Science
  • 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
  • In the past few years, convolutional neural nets (CNN) have shown incredible promise for learning visual representations. In this paper, we use CNNs for the task of predicting surface normals from a single image. But what is the right architecture? We propose to build upon the decades of hard work in 3D scene understanding to design a new CNN architecture for the task of surface normal estimation. We show that incorporating several constraints (man-made, Manhattan world) and meaningful… CONTINUE READING
    249 Citations

    Figures, Tables, and Topics from this paper

    Self-supervised Learning for Single View Depth and Surface Normal Estimation
    • 8
    • PDF
    Joint prediction of depths, normals and surface curvature from RGB images using CNNs
    • 22
    • PDF
    Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture
    • D. Eigen, R. Fergus
    • Computer Science
    • 2015 IEEE International Conference on Computer Vision (ICCV)
    • 2015
    • 1,556
    • Highly Influenced
    • PDF
    Marr Revisited: 2D-3D Alignment via Surface Normal Prediction
    • 173
    • PDF
    Face Normals "In-the-Wild" Using Fully Convolutional Networks
    • 31
    • Highly Influenced
    • PDF
    Deep Surface Normal Estimation on the 2-Sphere with Confidence Guided Semantic Attention
    • PDF
    Dense Optical Flow Prediction from a Static Image
    • 152
    • PDF
    Floors are Flat: Leveraging Semantics for Real-Time Surface Normal Prediction
    • 5
    • PDF
    Symmetry-aware Depth Estimation using Deep Neural Networks
    • 2
    • PDF

    References

    SHOWING 1-10 OF 46 REFERENCES
    Predicting Depth, Surface Normals and Semantic Labels with a Common Multi-scale Convolutional Architecture
    • D. Eigen, R. Fergus
    • Computer Science
    • 2015 IEEE International Conference on Computer Vision (ICCV)
    • 2015
    • 1,556
    • Highly Influential
    • PDF
    Depth Map Prediction from a Single Image using a Multi-Scale Deep Network
    • 1,757
    • PDF
    Rich Feature Hierarchies for Accurate Object Detection and Semantic Segmentation
    • 12,806
    • PDF
    DeepPose: Human Pose Estimation via Deep Neural Networks
    • 1,542
    • PDF
    ImageNet classification with deep convolutional neural networks
    • 59,531
    • Highly Influential
    • PDF
    A category-level 3-D object dataset: Putting the Kinect to work
    • 186
    • Highly Influential
    • PDF
    Learning Depth from Single Monocular Images
    • 839
    • PDF
    Structured Forests for Fast Edge Detection
    • 817
    • PDF