Monocular Differentiable Rendering for Self-Supervised 3D Object Detection

@inproceedings{Beker2020MonocularDR,
  title={Monocular Differentiable Rendering for Self-Supervised 3D Object Detection},
  author={Deniz Beker and H. Kato and Mihai Adrian Morariu and T. Ando and T. Matsuoka and Wadim Kehl and Adrien Gaidon},
  booktitle={ECCV},
  year={2020}
}
  • Deniz Beker, H. Kato, +4 authors Adrien Gaidon
  • Published in ECCV 2020
  • Computer Science
  • 3D object detection from monocular images is an ill-posed problem due to the projective entanglement of depth and scale. To overcome this ambiguity, we present a novel self-supervised method for textured 3D shape reconstruction and pose estimation of rigid objects with the help of strong shape priors and 2D instance masks. Our method predicts the 3D location and meshes of each object in an image using differentiable rendering and a self-supervised objective derived from a pretrained monocular… CONTINUE READING
    4 Citations

    Figures and Tables from this paper

    Curiosity-driven 3D Scene Structure from Single-image Self-supervision
    • PDF
    Vehicle Reconstruction and Texture Estimation Using Deep Implicit Semantic Template Mapping
    • PDF

    References

    SHOWING 1-10 OF 49 REFERENCES
    Autolabeling 3D Objects With Differentiable Rendering of SDF Shape Priors
    • 15
    • PDF
    Accurate Monocular 3D Object Detection via Color-Embedded 3D Reconstruction for Autonomous Driving
    • 44
    • PDF
    ROI-10D: Monocular Lifting of 2D Detection to 6D Pose and Metric Shape
    • 76
    • PDF
    3D-RCNN: Instance-Level 3D Object Reconstruction via Render-and-Compare
    • 131
    • PDF
    DirectShape: Photometric Alignment of Shape Priors for Visual Vehicle Pose and Shape Estimation
    • 12
    • PDF
    3D Packing for Self-Supervised Monocular Depth Estimation
    • 54
    • PDF
    MonoGRNet: A Geometric Reasoning Network for Monocular 3D Object Localization
    • 64
    • PDF
    DirectShape: Direct Photometric Alignment of Shape Priors for Visual Vehicle Pose and Shape Estimation
    • 5
    • PDF
    Single-Stage Monocular 3D Object Detection with Virtual Cameras
    • 2