In my special pet door that only allows cats to enter, I included 32 pictures of cats in the class 1 dataset, and another 32 pictures of dogs, lizards, and other animals in the class 2 dataset. So when a picture of a cat is shown, the machine (pet door) will detect class 1 and open the door. When a picture of a dog or another animal is shown, it will detect the class 2 dataset, keeping the door shut.  A limitation I encountered was a lack of data, primarily due to time constraints and limited resources, which prevented me from acquiring a real-life cat to include in dataset 1. This presents an issue because the machine relies on the 3D dimensions of a cat’s facial and bodily structure for accurate detection. Consequently, in some cases, the pet door may be misinterpreted a real-life cat, causing the door to remain closed. Overall, the machine did work the way I wanted. When I showed it a picture of a cat it leaned towards class 1, and when I showed it a dog or other animal it classifed it as class 2.