awesome-dynamic-slam

The following papers are the papers that fits my PhD research interest. They are mainly tackling the real-time reconstruction of multiplle dynamic objects or non-rigid objects. Some papers are off-line work yet also very interesting.

Due to my personal interests, SFM-based or factorization-based work are not collected here. Those papers can be found here.

Another related paper list is about the semantic segmentation, which can be found here.


Related survey papers:

  • State of the Art in Real-time Registration of RGB-D Images (link)
  • Visual SLAM and Structure from Motion in Dynamic Environments: A Survey (link)
  • State of the Art on 3D Reconstruction with RGB-D Cameras (link)

Multiple Rigid Object Tracking and Reconstruction (Real-time)

Non-rigid body Tracking and Reconstruction (Real-time)

Background only, i.e. removing dynamic objects

Non-realtime: non-rigid reconstruction

Non-realtime: multiple object segmentation-tracking-reconstruction

Learn Depth and camera motion

  • DeMoN: Depth and Motion Network for Learning Monocular Stereo

  • Unsupervised Learning of Depth and Ego-Motion from Video

  • Multi-view Supervision for Single-view Reconstructionvia Differentiable Ray Consistency

Learn 3D map

  • OctNet: Learning Deep 3D Representations at High Resolutions