This is a start to incorporate Microsoft Cognitive Services to provide a description or tags of what the camera can see. I am very happy with my initial testing. As I am living on the East Coast of Prince Edward Island, the weather has been too brutal to bring the camera outside. I am hoping to capture our wildlife in the backyard when winter behaves itself. We see an abundance of wildlife including foxes and crows. So for now my ceramic cat will do for the initial experiment.
After some trial and error I decided to go with Node Red that comes installed with the Raspberry Pi image. You need to upgrade to the most most current version.
This is my flow to snap a picture and send the description of the picture and the picture to Freeboard.
I have added a few nodes to make this work including dweetio, freeboard, Microsoft Cognitive Services , DropBox and RPI take photo. All of the nodes can be found at: https://flows.nodered.org/
The Dropbox node is set to a static name and when a picture is taken the original picture is updated to the new one.
I placed my wooded cat in front of the camera and received this image...interesting text is in reverse. Well this is not a piano sitting in front of a desk but my wooden cat with a sign wishing you "Happy Pi Day".