Skip to main content
. 2019 Oct 4;8:e48571. doi: 10.7554/eLife.48571

Figure 11. DeepFly3D graphical user interface (GUI) applied to with the Human3.6M dataset (Ionescu et al., 2014).

Figure 11.

To use the DeepFly3D GUI on any new dataset (Drosophila or otherwise), users can provide an initial small set of manual annotations. Using these annotations, the software calculates the epipolar geometry, performs camera calibration, and trains the 2D pose estimation deep network. A description of how to adopt DeepFly3D for new datasets can be found in the Materials and methods section and, in greater detail, online: https://github.com/NeLy-EPFL/DeepFly3D. This figure is licensed for academic use only and thus is not available under CC-BY and is exempt from the CC-BY 4.0 license.