I'm actually very surprised none of these trail camera companies are on board yet / have any AI detection tools implemented (I don't use any, so I could just be out of the loop and they do). I remember having the idea of utilizing the AI years ago before it became mainstream to detect animals via trail cams images while in school pursing a Comp. Sci degree. Models then for animal detection were actually largely done and trained on african safari animals (I believe), so the models to detect north american specific species were fairly limited (due to dataset training data) and were fairly off / incorrect at times. I couldn't see why models now days wouldn't be fairly great though and could help you sort and categorize your trail camera images by species and maybe even by sex.
Would be cool if a model could allow you to set an animal of interest, then you provide a few images of such animal, model be trained on that subset of images, and then have images be sorted via that named animal or even notified is said animal tripped a camera. Again, not anything simple to do though since you would really need hundreds if not thousands of images of said animal for it to be anything but accurate. Additional complexity comes in when you factor in antler growth, no one year let alone a week in growth would be consistent. Models would have to be trained daily to even have a chance at following and recognizing a specific target animal.
Definitely cool though no the less!
