Figure 1.

The original image (a) is mapped to a four-channel output (b), where the location of the left ear, right ear, shoulder, and tail are represented by Gaussian kernels in channels 1–4 of the output, respectively. Note that the four colors used in (b) are purely for visualization of the four separate channels of the target image. The overlay in (c) is provided to illustrate the locations relative to the original image.