Usually my first image will be the most “id-able” but I’m curious if the tool looks at all of them.
I believe it only uses the first.
Yes, the first. On the Android app, if you add three different organisms to a new observation and then go to suggestions, the suggestions will change as you swipe through each photo.
I also believe that to be the case. For that reason, if I’m uploading an observation that I’m uncertain about, I’ll sometimes get cv to look at each picture individually before merging them into a single observation. The level of consistency between the computer responses helps me to decide what level I’ll use for my id.
If you use the iNat Enhancement Suite – once the obs is uploaded, you can still ask CV to evaluate each picture. Especially useful when their sequence of photos is blurry first, then they get the subject in focus.