Skip to main content
. 2021 Sep 9;11(9):2655. doi: 10.3390/ani11092655

Figure 3.

Figure 3

Example images for annotations of the network in which the human observers creating high-quality annotations did not agree since a clear assessment was difficult due to shadows, lighting, view of camera, and/or the animals’ posture. The originally labeled annotations on the sample images are blue-colored and framed with red bounding boxes for easier identification.