Skip to main content
. 2020 Oct 8;14:51. doi: 10.3389/fnbot.2020.00051

Figure 7.

Figure 7

Single-grasping Annotations: (A) Good-grasping parallelogram annotations on original objects' event data; (B) Bad-grasping parallelogram annotations on original objects' event data. There are five typical objects selected, and there are four sampled time points in each recorded moving period. The first column presents the real grasping objects in traditional RGB image; the rest columns present the annotations and original objects' event data. The red edges present the grasping edge; and the green edges present the auxiliary edges.