The practice of capturing or documenting animals through visual images reveals much about the relationship between humans and animals. The highly informative photographs of animals found in common nature documentaries or on the internet are deeply imbued with the visual grammar through which humans view wildlife. In contrast, images captured by trail cameras (unmanned cameras), which lack human intervention, reveal a broader array of perspectives. These cameras do not (and cannot) capture animals in idealized poses like human photographers might. By compiling these unmanned camera images into datasets and conducting machine training, the resulting representations of animals are often blurred but typically appear as side or frontal views, contrary to the varied orientations in the dataset. These forms inadvertently reveal the visual systems humans use when photographing animals. Meanwhile, the blurred, ghost-like stains that form no clear shape feel paradoxically like the true portraits of these beings.