1. On the ImageNet webpage, it says that it does not own the image and it just compiled a bunch of images on the web for it to use. I do wonder where specifically they get them from, like if it was ripped from Google Images or something. I noticed that some of the photos were of people doing certain activities, or shots of people at certain occasions. Would it know if the image is okay to use?
2. I used the Image Classifier with webcam and it wasn’t really able to get many things with a high confidence or really close at all. I showed it things like pens, my water bottle, phone, wallet, glasses, charging cables. hairbands, and the closest I felt it was was calling a phillips head screwdriver a drill. It for some reason also liked to default to “camouflage”, which may be because my head was in frame? I also feel like when I tried it the fact that my camera was facing out the window may have also been a factor.