Revealing Occlusions with 4D Neural Fields

By Basile Van Hoorick, Purva Tendulkar, D'idac Sur'is, Dennis Park, Simon Stent, Carl Vondrick. 2022

24 citations recorded by the source corpus.

For computer vision systems to operate in dynamic situations, they need to be able to represent and reason about object permanence. We introduce a framework for learning to estimate 4D visual representations from monocular RGB-D video, which is able to persist objects, even once they become obstructed by occlusions. Unlike traditional video representations, we encode point clouds into a continuous representation, which permits the model to attend across the spatiotemporal context to resolve occlusions. On two large video datasets that we release along with this paper, our experiments show that the representation is able to successfully reveal occlusions for several tasks, without any architectural changes. Visualizations show that the attention mechanism automatically learns to follow occluded objects. Since our approach can be trained end-to-end and is easily adaptable, we believe it will be useful for handling occlusions in many video understanding tasks. Data, code, and models are available at occ1usions. cs. co1umbia. edu.

Subfields: Computer Vision and Pattern Recognition, 3D Vision & Pose

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Computer Vision and Pattern Recognition

  1. Pavel Tokmakov
  2. Sergey Zakharov
  3. Rundi Wu
  4. Basile Van Hoorick
  5. Carl Vondrick
  6. Ruoshi Liu
  7. Luc Van Gool
  8. Andrew Zisserman

In 3D Vision & Pose

  1. Rundi Wu
  2. Carl Vondrick

Citations in the corpus