I’m trying to create an automation where after uploading a photo to Google Drive, Amazon Rekognition will check and assign the so-called labels to the photo, i.e. identify if there are any objects in the photo and then name them.
And I have a problem because in the output from Rekognition I see that I have about 10 different types of labels. For example, one that defines the photo as an advertisement, which is OK (screenshot).
However, when I want to send a message to Slack and try to map it, I don’t see these different types of outputs at all.
There are many of them in Rekognition output. What exactly should I choose here for a message to be sent to Slack that will contain a list of labels assigned by Rekognition to the photo?
I am attaching blueprint
blueprint (4).json (46.7 KB)
The Labels object in the Rekognition output is an array.
Slack doesn’t handle arrays - only text. So you need to process that array object into a suitably formatted text list.
You’ll want to use an Iterator module with the Labels array as input, followed by a Text Aggregator module to aggregate your selected items from the array (I’m guessing just the Name parameter).
The free Make Academy has a module on “An introduction to aggregators” in the Basics course.
I’d recommend that you set aside some time for the entire Basics course - and ideally the Intermediate course as well.
It’s well worth the short time investment!
2 Likes
Thank you. I managed to finish label part with the help of ChatGPT but I am actually going through the academy. But the outcome is exactly as I want.
Now I need to work on text detection. I think if it makes sense to show confidence or just detected text, which is moch easier task.
Thank you
1 Like
