Semantic Foreground Inpainting From Weak Supervision

  title={Semantic Foreground Inpainting From Weak Supervision},
  author={Chenyang Lu and Gijs Dubbelman},
  journal={IEEE Robotics and Automation Letters},
  • Chenyang Lu, Gijs Dubbelman
  • Published 2020
  • Computer Science
  • IEEE Robotics and Automation Letters
  • Semantic scene understanding is an essential task for self-driving vehicles and mobile robots. In our work, we aim to estimate a semantic segmentation map, in which the foreground objects are removed and semantically inpainted with background classes, from a single RGB image. This semantic foreground inpainting task is performed by a single-stage convolutional neural network (CNN) that contains our novel max-pooling as inpainting (MPI) module, which is trained with weak supervision, i.e., it… CONTINUE READING

    Figures, Tables, and Topics from this paper.


    Publications referenced by this paper.
    Context Encoders: Feature Learning by Inpainting
    • 2,021
    • PDF
    Fully Convolutional Networks for Semantic Segmentation
    • 7,336
    • PDF
    The Cityscapes Dataset for Semantic Urban Scene Understanding
    • 3,293
    • Highly Influential
    • PDF
    SegNet: A Deep Convolutional Encoder-Decoder Architecture for Image Segmentation
    • 3,622
    • Highly Influential
    • PDF
    Very Deep Convolutional Networks for Large-Scale Image Recognition
    • 40,144
    • PDF
    Conditional Random Fields as Recurrent Neural Networks
    • 1,854
    • PDF
    Pyramid Scene Parsing Network
    • 2,896
    • Highly Influential
    • PDF