Date: January 2020.
Source: International Journal of Computer Vision 128, 873–890 (2020). https://doi.org/10.1007/s11263-019-01279-w.
Abstract: The optical flow of humans is well known to be useful for the analysis of human action. Recent optical flow methods focus on training deep networks to approach the problem. However, the training data used by them does not cover the domain of human motion. Therefore, we develop a dataset of multi-human optical flow and train optical flow networks on this dataset. We use a 3D model of the human body and motion capture data to synthesize realistic flow fields in both single- and multi-person images. We then train optical flow networks to estimate human flow fields from pairs of images. We demonstrate that our trained networks are more accurate than a wide range of top methods on held-out test data and that they can generalize well to real image sequences. The code, trained models and the dataset are available for research.
Article: Learning Multi-human Optical Flow.
Authors: Anurag Ranjan, David T Hoffmann, Dimitrios Tzionas, Siyu Tang, Javier Romero, Michael J Black. Max Planck Institute for Intelligent Systems, Tübingen, Germany.

Related Articles:

  1. Capture, Learning, and Synthesis of 3D Speaking Styles. D Cudeiro, T Bolkart, C Laidlaw, A Ranjan, MJ Black.
  2. Learning to Dress 3D People in Generative Clothing. Q Ma, J Yang, A Ranjan, S Pujades, G Pons-Moll, S Tang, MJ Black.
  3. Expressive Body Capture: 3D Hands, Face, and Body from a Single Image. G Pavlakos, V Choutas, N Ghorbani, T Bolkart, AA Osman, D Tzionas, MJ Black.