I understand that object recognition is possible through the Amazon Rekognition add-on. But can I point to my own model (hosted on Amazon Rekognition) instead of getting the default objects recognized?
Using custom models is not supported at this moment, but I think it’s something that we may consider for future updates. I’m moving this to #feature-requests
Thanks for sharing your idea with us! I can’t give any estimations at this point, but we’ll be tracking demand for this functionality and will be in touch if we have any news.