If only I could figure out exactly how to do them/what they want.. Do they want you to do sentences to label everything in every photo or what?
You are supposed to box the objects listed in the phrase provided. The instructions seem pretty clear to me, but I've done quite a few VGs so maybe I "get" them a little better. Have you read the examples? If they show a phrase reading "Cat on a table", you enter "cat" as the object, draw a box around the cat, then enter "table" as the object, and draw a box around the table. If it says "large brown dog", then you just enter "dog", and draw a box around the dog. The blue boxes are there to tell you what part of the picture the phrase is referring to, in case there are, say, multiple dogs in the picture.