Learning Human Optical Flow
This work addresses the need for efficient and accurate human optical flow estimation, which is useful for human action analysis, but it is incremental as it builds on existing methods like SpyNet.
The authors tackled the problem of estimating optical flow specifically for human motion by creating a new training database using 3D models and motion capture data, and training a convolutional neural network based on SpyNet. They demonstrated that their method is more accurate than top generic flow methods on test data and generalizes well to real images, with code and dataset made available.
The optical flow of humans is well known to be useful for the analysis of human action. Given this, we devise an optical flow algorithm specifically for human motion and show that it is superior to generic flow methods. Designing a method by hand is impractical, so we develop a new training database of image sequences with ground truth optical flow. For this we use a 3D model of the human body and motion capture data to synthesize realistic flow fields. We then train a convolutional neural network to estimate human flow fields from pairs of images. Since many applications in human motion analysis depend on speed, and we anticipate mobile applications, we base our method on SpyNet with several modifications. We demonstrate that our trained network is more accurate than a wide range of top methods on held-out test data and that it generalizes well to real image sequences. When combined with a person detector/tracker, the approach provides a full solution to the problem of 2D human flow estimation. Both the code and the dataset are available for research.