Skip to main content
. 2019 Mar 6;9:3744. doi: 10.1038/s41598-019-40064-0

Figure 5.

Figure 5

(a) Depth map from the developed setup (raw data, no post-processing). (b) Conventional Image of scene for display purposes. (c) Depth map from the Kinect used as reference. The yellow square corresponds to the field of view. (d) Relative error for this scene, with the output of a Microsoft Kinect as ground truth. A sparse set of handpicked points were selected in the ground truth and then compared to depth estimations from our network.